CHAPTER Xlll THE MARKET BREAK OF SUMMARY TABLI~ OF CONTE~TS A. Introduction. B. Background. C. The Days of the Break. D. Activities of the Participants. E. Summary, Conclusions, and Recommendations.

815 DETAILED TABLE OF CONTENTS

A. INTRODUCTION

1. The significance of the break ...... 821 2. Scope of the chapter ...... 822 3. Method of study and sources of data ...... 823

B. BACKGROUND 1. The economic framework ...... 824 2. The securities markets ...... 824 3. Investment advice ...... 830 4. The immediate prelude--The week before the break ...... 831 C. THE DAYS OF THE BREAK 1. Monday, , 1962 ...... 832 2. Tuesday, , 1962 ...... 834 3. Wednesday, , 1962 ...... 835 4. Thursday, , 1962 ...... 835 5. Reaction of foreign markets ...... 836 6. Epilogue ...... 836 D. ACTIVITIES OF THE PARTICIPANTS 1. Nonmembers of the NYSE ...... 838 a. Public individuals ...... 838 (1) Aggregate activity ...... 838 (2) Activity in the eight stocks ...... 839 (a) American Telephone & Telegraph Co ...... 839 (b) Avco Corp ...... 840 (c) Brunswick Corp ...... 840 (d) General Motors Corp ...... 840 (e) International Business Machines Corp ...... 840 (f) E.J. Korvette, Inc ...... 841 (g) Standard Oil Co. (New Jersey) ...... 841 (h) Steel Corp ...... 841 b. Foreign participation...... 841 (i) Aggregateactivity ...... 841 (2) Activityin the eight stocks ...... 843 c. Institutionalparticipation ...... 844 (i) Aggregateactivity ...... 844 (2) Activityin the eight stocks...... 844 d. Open-end investment companies ...... 845 (1) Aggregateactivity ...... 845 (2) Activityin the eight stocks...... 845 (a) American Telephone & Telegraph Co ...... 846 (b) General Motors Corp ...... 846 (c) InternationalBusiness Machines Corp ...... 846 (d) StandardOil Co. (New Jersey)...... 847 (e) United States Steel Corp ...... 847 (f) The remainingthree stocks ...... 847 817 818 DETAILED TABLE OF CONTENTS

~Page 2. Members of ~he NYSE ...... 847 a. Members as a group ...... 847 b. Specialists ...... 848 (1) Aggregate activity ...... 848 (2) Activity in the eight stocks ...... 849 (a) American Telephone & Telegraph Co ...... 849 (b) Aveo Corp ...... 849 (e) Brunswick Corp ...... 850 (d) General Motors Corp ...... 850 (e) International Business Machines Corp ...... 850 (f) E. J. Korvette, Ine ...... 851 (g) Standard Oil Co. (New Jersey) ...... 851 (h) United States Steel Corp ...... 851 e. Floor traders ...... 851 (1) Aggregate activity ...... 852 (2) Ab~tivity in the eight stocks ...... 852 (a) American Telephone & Telegraph Co ...... 853 (b) Aveo Corp ...... 853 (c) Brunswick Corp ...... 853 (d) International Business Machines Corp ...... 853 (e) Standard Oil Co. (New Jersey) ...... 853 d. Members off-floor ...... 854 (1) Aggregate activity ...... 854 (2) Activity in the eight stocks ...... 855 (a) American Telephone & Telegraph Co ...... 855 (b) Aveo Corp ...... 1855 (e) General Motors Corp ...... ~855 (d) E. J. Korvette, Ine ...... 855 e. Odd-lot dealers ...... 855 (1) Aggregate activity ...... 856 (2) Activity in the eight stocks ...... 857 (a) American Telephone & Telegraph Co ...... 857 (b) International Business Machines Corp ...... 857 3. The participants in perspective ...... 857 E. SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS CHA R TS Chart No. Part Page a. Dow-Jones Industrial Average and NYSE volume, to , 1962, on a weekly basis ...... B 827 b. Dow-Jones Industrial Average, April 2 through July 31, 1962, daily high, low, close ...... B 828 c. American Stock Exchange, New York Herald Tribune average of 25 stocks, September 1961 to June 30, 1962, on a weekly basis ...... B 820 d. National Quotation Bureau industrial average of 35 over-the- counter stocks, September 1961 to June 30, 1962, on a weekly basis ...... B 829 e. Weekly member and nonmember groups’ purchase and sale bal- ances from stock transactions on the New York Stock Ex- change ...... D facing 838 f. Cumulative purchase and sale balances of domestic stocks by foreigners ...... D 842 (At end of chapter) 1. American Telephone & Telegraph Co.: Price movement, volume, and members’ transactions May 28, 1962 ...... facing 864 2. Avco Corp.: Price movement, volume, and members’ transactions May 28, 1962 ...... facing 864 3. Brunswick Corp.: Price movement, volume, and members’ transactions May 28, 1962 ...... facing 864 4. General Motors Corp.: Price movement, volume, and members’ trans- actions May 28, 1962 ...... facing 864 5. International Business Machine Corp.: Price movement, volume, and members’ transactions May 28, 1962 ...... facing 864 6. E. J. Korvette, Inc.: Price movement, volume, and members’ trans- actions May 28, 1962 ...... facing 864 DETAILED TABLE OF CONTENTS 819

Chart No. Page 7. Standard Oil Co. (New Jersey): Price movement, volume, and mem- b~’rs’ transactions May 28, 1962 ...... facing 864 8. United States Steel Corp.: Price movement, volume, and members’ transactions May 28, 1962 ...... facing 864 TABLES Table No. Part Page a. Sales of General Motors Corp. emanating from the , selected days, 1962 ...... _~ ..... D 844 b. Rankings of five stocks held by funds, selected periods, 1961- D 846 c. Ratio of floor traders’ purchases and sales to total volume, May 28, 1962 D 852 d. Distribution of types of orders on the NYSE, May 28, 29, and E 860 (At end of chapter) 1. Comparison of stock prices on the NYSE, Amex, and over the counter (selected days, 1961-62) ...... 865 2(a-h). Round-lot purchases and sales of eight selected stocks by mem- bers and nonmembers (19 selected days, 1961-62) ...... 866 3(a-h). Round-lot purchases and sales of eight selected stocks by non- members classified by type of nonmember (19 selected days, 1961- 62) ...... : ...... 874 4. Percent distribution of shares purchased and sold by public individuals in eight selected stocks, by type of order (May 28, 29, and 31, 1962)_ 882 5. Round-lot trading in eight selected stocks by public individuals in cash and margin accounts (May 28, 29, and 31, 1962) ...... 883 6. Odd-lot trading in eight selected stocks by public individuals in cash and margin accounts (May 28, 29, and 31, 1962) ...... 884 7. Purchases and sales of U.S. stocks by foreigners (1950-62) ...... 885 8. Common-stock trading on the NYSEby selected institutions classified by type of institution (monthly, September 1961 to ) .... 886 9. Common-stock trading on the NYSEby selected institutions classified by type of institution (weekly, April to June 1962 and daily, May 28 to 31, 1962) ...... 887 10(a-e). Round-lot trading in five selected stocks by selected institutions (September 1961 to June 1962) ...... 889 11. Common-stock trading on the NYSE by selected open-end investment companies (monthly, September 1961 to June 1962 and weekly, April to June 1962) ...... 894 12(a-h). Round-lot trading in eight selected stocks by selected open-end investment companies (September 1961 to June 1962) ...... 895 13. Percent distribution of shares purchased and sold in round lots by members in eight selected stocks, by type of member (19 selected days, 1961-62) ...... 903 14(a-h). Round-lot purchases and sales of eight selected stocks by mem- bers classified by type of member (19 selected days, 1961-62) ...... 904 15. Specialists’ activity, market movements and total volume on the NYSE (weekly, January to June 1962) ...... 920 16. Round-lot trading on the NYSE by members classified by type of member (May 28, 29, and 31, 1962) ...... 920 17. Floor traders’ activity, market movements, and total volume on the NYSE (weekly, January to June 1962) ...... 921 18. Floor traders’ total sales and short sales on the NYSE (weekly, January to June 1962) ...... 921 19. Round-lot and odd-lot trading by odd-lot dealers on the NYSE (weekly, January to June 1962) ...... 922 APPENDIX CHARTS

APPENDIX B Chart No. l(a-h). Price movement and volume, eight selected stocks, September 1961 to June 30, 1962, and summary of investment advice ..... 927 820 DETAILED TABLE OF CONTENTS

APPENDIX TABLES APPENDIX B Table No. Page l(a-h). Volume and prices of eight selected stocks (19 selected days, 1961-62) ...... 931 2(a-h). Volume and prices of eight selected stocks (monthly, September 1961 to June 1962, and weekly, April to June 1962) ...... 935 APPENDIXES A. Methodology ...... 923 [ B. Price movements and volume in eight selected stocks ...... 927 C. Letters to NYSEclearing firms and to firms in their omnibus accounts concerning transactions on 16 specific dates and for eight selected stoc 939 D. Questionniares IN-l, IN-2, and IN-3: Inst~itutional investors’ trans- actions in common stocks ...... 943 E. Questionnaires EX-5 and EX-6: NYSE members’ and member organi- zations’ off-floor transactions for eight selected stocks ...... 951 A. INTRODUCTION

1. THE SIGNIFICANCE OF THE BREAK The last 3 trading days of May 1962 witnessed unique and un- expected events in the securities markets of the world. On Monday, May 28, on the New York Stock Exchange (NYSE) the Dow-Jones Industrial Average on very heavy volume dropped 35 points, the sec- ond largest decline in points ever registered by the average. On May 29 the average rose 27 points. Reported volume was 14=,750,000 shares, second only to the 16 million shares traded on October 29, 1929. On May. 31, after the MemorialDay holiday, the average rose nine points, again on heavy trading. Similar severe price fluctuations occurred on the American Stock Exchange (Amex), the regional exchanges and in the over-the-counter market. These events, especially thoze on the NYSE, have come to be known as the "market break of May 1962." The markets’ erratic behavior prompted concern and caused be- wilderment at homeand abroad. ’The frenetic activity of the break reJ sulted in large and sudden losses for many and gains for some. There was concern in Government and business circles that this break, like previous breaks of similar magnitude, might signal or provoke a seri- ous business recession. There were unconfirmed but active rumors that "professional speculators" had deliberately triggered the decline. Some corporations reconsidered their plans for capita] expenditures and a number of previously scheduled stock offerings were postponed or canceled. Although more than half of the pe(~ple living in the United States were born after the crash of 19"29, the memoryof that event still casts a shadow over every major market drop, and, there- fore, this break had a strong and immediate psychological impact upon the Nation. Like previous breaks in stock market history, the break of May1962 was the occasion for a deluge of retrospective analysis. There was little agreement as to its causes and even less as to the possible con- sequences. Some said it signaled the "end of inflation as a way of life" and others called it the "definitive conclusion" to the longest bull market in history. On the other hand, others saw it only as a "temporary correction" in the continuing upward trend of stock prices. There were also contradictory analyses of the roles played by the key participants. For example, some said specialists were buying; others said they were selling. These contradictory analyses, as well as the ma~o~nitude of the de- cline on May 28 and recovery on May 29 and 31, clearly indicated .the need for a detailed and careful analysis of the events and their causes. In an effort to determine more precisely just what happened both the New York Stock Exchange and the Securities and Exchange Commission announced studies of the break. ’Chairman William L. 821 822 REPORT OF SPECIAL STUDY OF SECURITIES MARKETS Cary speaking before the House Interstate and Foreign Commerce Committee on May 29 said : We should be aware of the question * * * of who or what groups are buying and selling and the status of specialist accounts and the like. * * * We are look- ing1 into these facts. Later in August he elaborated in these words: ~Te have decided that it was necessary to make an analysis of the whole situa- tion of who were buying and who were selling during that period around May 28. We are doing a broad analysis * * * and we will hopefully secure a composite picturee of the market at that time, of the buying and selling, It was decided that this analysis of the market break should be under- taken by the Special Study of Securities Markets. What follows is the result of this examination.

2. SCOPE OF THE CI~AI~ER The entire securities industry was directly affected by the break and certain market practices were undoubtedly affected by or con- tributed to its severity. As a result~ muchof the analysis of the break appears in preceding chapters of this report. For example~ the con- duct of specialists during the break is treated extensively in chapter VI~ the reaction of the over-the-counter market to the break is dis- cussed in chapter VII~ and margin selling is discussed in chapter X. Accordingly~ one purpose of ~his chapter is to combine these and additional materials into a unified analysis of the markets in a period of stress and draw such conclusions as may be applicable from such an appraisal. Since it deals with the same events~ this chapter covers in certain respects the same materials presented by the New York Stock Ex- change in its study of the market break~ "The Stock Market Under Stress~" issued in complete form in . How~ver~ there are several basic differences. One particularly fundamental differ- ence concerns the analysis of the statistical data. The NYSEin its analysis of such mutters as who was buying and selling~ the char- acteristics of the investors participating in the market~ and the activity of the Exchange~s member organizations during the break relied essentially upon ,aggregated dat~. On the other hand~ the Special Study~ in this area as in several others~ found that aggregated and averaged data~ although of unquestionable importance and usefulness for many purposes~ provided an insufficient picture of what occurred and in some ways might even be misleading2 The~fore~ unlike the NYSEstudy~ this chapter is concerned with particular trading in certain stocks ~nd where possible with individual transactions. It also deals with the n~ture of information available to investors and others during and immediately after the break. Following the summary of the sources of information is a short account of the background in which the break actually occurred~ including a brief description of the economic framework~ the con-

1Hearings on S. 2135 ("SEC Reorganization") before the House Committee on Inter- state and Foreign Commerce, 87th Cong., 2d sess., p. 19 (19.62). ~ Hearings before the sutmommittee of the Senate Committee on Appropriations on YI.R. 12711, 87th Cong., 2~1 sess., p. 1317 (1962). a Transmittal letter of Special Study., H. Doc. 95, pt. 1, 88th Cong., 1st sess., p. xvi (1963). REPORT OF SPECIAL STUDY OF SECURITIES MARKETS 823 dition of the securities markets, and the character of the investment advice being given. A brief summary of overall buying and selling is presented with reference to the NYSEstudy. Thereafter the chapter concentrates on presenting data not previously given with respect to the transactions in particular securities and, for members of the NYSE,at particular points during the trading days.

3. METHOD OF STUDY AND SOURCES OF DATA The opening, more general, sections dealing with the economic and market background are based upon a broad range of source mate- rials including selected interviews~ newspapers, economic and finan- cial journals, Governmentpublications, and standard reference works. The overall description of the days of the break relies in good part on the N¥SE research report, "The Stock Market Under Stress." The analysis of the activities of the participants in the market for the period studied is based largely on a detailed study of the trading in eight selected NYSEstocks. In addition, the longer term trend of trading in all stocks by various groups of participants has been analyzed on the basis of monthly and weekly data from September 1961 through June 1962. These data for NYSE member groups and for nonmembers in the aggregate are furnished to the Commis- sion by the Exchange each week and show daily round-lot purchases, sales, and short sales. The Commission regularly compiles figures on odd-lot trading from weekly reports of daily transactions filed by the odd-lot dealers2 The eight selected stocks were ones with high volumes before, dur- ing, .and after the break which also displayed varying patterns of price activity. Some of these are so-called "market leaders" and, as such, are often watched by investors as market barometers. More- over, four of the eight are included in the Dow-Jones Industrial Average, so that their price movements have a direct bearing on the general market price level (as popularly measured by the average) anygiven time. The commonstocks of the following companies were selected : American Telephone & Telegraph E. J. Korvette, Inc. Co. International Business Machines Avco Corp. Corp. Brunswick Corp. Standard Oil Co. (New Jersey) General Motors Corp. United States Steel Corp. Trading in these stocks was studied on 14 scattered days preceding the break; on the 3 days of the break; and on 2 days in June, after the break2 These detailed data made it possible to visualize the nature of activity prior to the break, as well as during the break itself, in each of the stocks studie~i. The transactions of nonmembersin the 8 stocks studied during the 19 days were obtained from data provided by the 25 largest NYSE member firms, measured by gross commission income. Additional nonmemberdata were obtained through Special Study questionnaires 4 See app. XIII-A for a more detailed description of the data. ~’rhes.e rotlnd-lot ,nd odd-lot d,ata are published weekly in an SEC Statistical Release and monthly in the SEC Statistical Bulletin. ~ The days studied were : Nov. 3, 6, 15, 16, and 17 in 1961 ; and in 1962, Jan. 29 an@ 30, Mar. 15 and 16, Apr. 27 and 30, , 11, 14, 28, 29, and 31, and. and 15. 824 REPORT OF SPECIAL STUDY OF SECURITIES MARKETS t.o 162 financial institutions and 55 open-end investment companies. These institutions submitted information concerning their activities in the eight stocks as well as some aggregate data. Supplemental statis- tics on foreign transactions were obtained directly from selected NYSE firms specializing in such accounts and from certain commercial banks. The transactions of the different categories of NYSEmembers were derived from regularly compiled reports, special requests, and ques- tionnaires. In addition to information on the eight stocks for the 19 days, data concerning the intraday activities in these stocks on May 28 and May 29 were obtained for the members (excluding mem- bers off-floor on May29), and charts presenting these transactions on a timed basis were constructed for May28.

B. BACKGROUND

1. THE ECONO/~[IC FRAMEWORK For the economyas a whole, the year and a half preceding the mar- ket break was a period of hesitation. The national economy’s recov- ery from the cyclical low of was disappointing to many economists. The basic trend of economic growth, while still upward, lagged behind expectations. Gains in employment, personal income, and industrial production were behind those attained at similar points in previous cyclical recoveries. General business activity in the early months of 1962 was slowing down. The percentage 1’ate of increase in the gross national product, generally considered the most inclusive measure of trends in the eco~- omyas a whole, was 0.9 percent and 0.7 percent for the first two qu~ r- ters of 1962. In terms of dollars it fell almost $30 billion below the anticipated~ figures. Another measure of the economy’s performance is contained in sta- tistics known as the Census Bureau’s "leading indicators," a series relating to bank debits, durable goods sales, business failures, employ- ment, and stock prices. Most of these statistics, which are watched for their alleged predictive ability, were faltering in the first half of 1962. The number moving to higher ground ~vas steadily dwindling. Thus the market break of May 1962 occurred within the frame~vo.rk of an economythat showed signs of uncertainty.

2. THE SEgJ~URITIF~ ~ARKI~TS Doubts concerning the economy were certainly not reflected in the exuberant stock markets of 1961. As noted in chapter IV, there was unusually high activity in new issues accompanied by an atmosphere of feverish speculation during the early months of 19612 The new 1961 offerings represented in many instances young, untried, small businesses frequently ~vith scientific-sounding namesending in -namics, -onics, or -mation. Among these were: Digitronics, Hedtronics, Pacotronics, Microsonics, Nucleonics, Techmation, Pneumodynamics. There is little doubt that some of the many first-time investors who eagerly bid for shares in these companies had little understanding of what these companies manufactured.

Economic Report of the President, transmitted to Cgngress , p. 7. Ch. IV.B.3. REPORT OF SPECIAL STUDY OF SECURITIES MARKETS 825

Not only was 1961 distinguished by a hyperactive new issue market, but it also was one of the busiest years for block distributions--the sales of large blocks of already outstanding securities by individuals, estates, or institutions. Reaching record levels in 1961, when 1960 share and dollar totals were doubled, block distributions further con- tributed to the process of saturating the public with shares of common stock. Pr!ce-ea.rnings ratios touched levels in 1961 attained but on few occasions m the past. This ratio, the relationship of" market price to annual earnings per share, has been generally considered an im- portant guide to stock value and a useful indication of comparative price levels. In 1961, investors apparently were willing to pay in- creasingly more for earnings. This is evident from a review of the rising price-earnings ratio of a large cross section of NYSEstocks as represented by the Standard & Poor’s "500" Stock !ndex and the 30 Dow-Jones Industrials, two %verages" commonly followed as guides to the action of the market as u whole. While the general level of earnings for the stocks comprising these indexes was rising gradu- ally, the price level rose even more rapidly. In the third quarter of 1961, the price-earnings ratio of the Standard & Poor’s "500" Stock Index reached 23.7 and the Dow-Jones Industrials 9_4.2. Including 1961, the 5-year (1957-61) average of the price-earnings ratio Standard & Poor’s Index was 17.9 times. A~c the end of 1961 in its "Annual Forecast--1962," Standard & Poor’s weekly advisory pub- lieation "The Outlook" commented on these higher price-earnings ratios: We believe that historic high appraisals must now be accepted. They are the product of a broadening base of stock ownership and the tendency, on the part of both the professional and the general investor, to seek long-term values. This does not mean an unending rise in valuations, but a new 16-20 plateau of price-earnings ratios, contrasting with the old norm of 10-13. The price-earnings ratios of certain "growth" and "glamor" stocks rose to even more exceptional levels. One spectacular and well-pub- licized example was Polaroid, the photography equipment company, which rose from an average price of 4 in 1953, on earnings of $0.37~ to 238 in 1961, on earnings of $2.07, having touched an alltime high of 261 the year before. Actually, Polaroid’s earnings were in a de- clining trend from 1959 through 1961, yet the price-earnings ratio was 100 in 1961. Other stocks with particularly high price-earnings ratios in 1961 were: International Business Machines~ 66; Texas In- struments, 66 ; Varian, 73 ; and Xerox, 76. The active merchandising techniques of the greatly expanded se- curities industry as discussed in chapters I and III undoubtedly con- tributed to the general price rises, high volumes~and high pric~-earn- ings ratios of 1961 (table XIII-1). ~ These generally rising averages, however, concealed the fact that actually many individual stocks and industry groups had been in their own private "bear" markets for several years. For example, aluminums and domestic oils had reached their peaks in 1957, metal and glass containers in 1958, rails and steels in 1959, and radio-TV broadcasters in 1960. By the end of 1961 a majority of the industry groups comprising both the SEC and the Standard & Poor’s Indexes had "topped out." Obscured too by the rising averages was the growing number of stocks declining and the smaller number of stocks advancing. The ratio of daily advances to declines had been steadily diminishing since early August. A similar 826 REPORT OF SPECIAL STUDY OF SECURITIES MARKETS declining ratio in the number of stocks reaching new highs to those reaching new lows was also observable toward the end of the year. This was accompanied by a declining overall volume on the ex- changes and especially over the counter. In contrast, there was ~ renewed pickup and interest in other investment media. Beginning in 1962 there was an unusual and marked rise in savings bank de- posits. Contributing to this rise was the authorization of higher in- terest rates by the Federal Reserve System and the Federal Deposit Insurance Corporation in . Resulting bank demand also sparked renewed interest in the corporate and municipal bond markets, both of which showed their first real gains in some 15 years, At least two signs of "professional" disenchantment with the securi- ties markets preceded the break. (1) Open-end investment com- panies between January and increased their liquid assets some 25 percent, from $980 million to $1,284 million. 9 However, there was at that time little unanimity among the fund managers as to the f’uture course of the market. According to one survey, some were con- sidered "quietly bearish" while others saw some strength left in the m~rket. Most agreed that the high price-earnings ratios were influenc- ing their decisions in selecting securities. 1° (2) The price of a seat the New York Stock Exchange reached a 28-year peak of $225,000 on March 24, 1961, and again on August 8, 1961, and declined thereafter at a rapid rate--to $200,000 on January 17, 1962; $175,000 on March 27; $160,000 on April 18; and $150,000 on . These transac- tions gave an indication of what estimates the purchasers placed on the~1 prospects ahead. Although there were some intermittent rallies in the level of the averages in early 1962, none of these rallies carried prices to their recent historic highs. After March, the market levels on the NYSE declined at a rapid pace. Through most of April and May the rate of decline accelerated on increasing volume (charts XIII-a and XIII-b). The American Stock Exchange Index reached its peak in November1961, and while the decline paralleled that of the Dow-Jones Industrials, the relative price drop was somewhat greater (chart XIII-c). The events on the over-the-counter market paralleled those on the NYSE and Amex. From the beginning of 1962, many over-the- counter dealers reduced the size of their markets as interest in over- the-counter securities declined and their own views of the future grew more bearish. :Naturally, not all firms shared a pessimistic view of m~rket prospects but a majority apparently were inclined to reduce their inventories and some discontinued tradin.g in certain more spec- ulative issues. The only index measuring price change in the over- the-counter market is compiled by the Na.tion~l Quotation Bureau and may not be representative of the variety of securities traded in this market. 1~ The NQBAverage followed the averages of exchange listed stocks downwardin early 1962 (chart XIII-d).

~ Includes cash, U.S. Government and short-term bonds. Source: Investment Company Institute. ~o Business Week. Feb. 17, 1962, p. 110. ~ On . 1962, the first seat transfer following the break took place at $115.000, the low point for the year. Thirteen d~ays later the Dow-Jones Inc~ustrials close(~ at the year’s low. Subsequently seat prices moved up along with the averages for the rest of 1962. ~ This avesage is currently made up of 35 issues such as American Express, Anheuser- Busch, Inc., Eli Lilly & Co., tTime, Inc., and ~reyerhaeuser, and is not generally represent- ative of the lower priced and more s~eculative t~’De ~f -~urity which has been extensively traded in this market. REPORT OF SPECIAL STUDY OF SECURITIES MARKETS 827

ChartXlll-a DOW-JONES INDUSTRIAL AVERAGE ~nd N.Y.S.E. VOLUNE September 1961 to June 30, 1962 on a weekly basis

NYSE VOLUME SHARES MILLIONS -- ,~0

-- 30

NY.S.E. VOLUME - 20

- 10

_ 0 J I F I M I A I M 961 ]962

96-746 O---63--pt. 4-----~4 828 REPORT OF SPECIAL STUDY OF SECURITIES mARKETS

Chart Xlll-b DOW-JONESINDUSTRIAL AVERAGE tDaily high, low, close)

April 2 through July 31, 1962 700

620

60O ,I

58O

560

540

520 III1111111111111111IIIIIIIIIIIIIIIIIIII IIIIIIIIIIIIIIIII1 April May June

4448 REPORT OF SPECIAL STUDY OF SECURITIES ~RKETS 829

ChartXIII-c AMERICAN STOCK EXCHANGE PRICE AVERAGE* and VOLUME September 1961 to June 30, 1962 on a weekly basis AVERAGE 160 --

150 --

140 -- AVERAGE* VOLUME (Close) SHARES MILLIONS I~0 -- - 12

120 -- VOLUME

II0 --

100

0 I 1961 1962 New York Herald Tribune - Average of 25 stocks.

Chart Xlll-d NATIONAL QUOTATION gUREAU INDUSTRIAL AVERAGE of 35 Over-The-Counter Stocks AVERAGE September1961 to June 30, 1962 on a weekly basis 160

150

140

AVERAGE 130 (Close)

120

110

j I F I M I A I M i J 1961 1962 830 REPORT OF SPECIAL STUDY OF SECURITIES ~¢IARKETS

3. INVESTlVIENT ADVICE The record of investment advice over this period bears out what seems to be a commontendency to interpret the future in terms of the immediate past. Following the declines of 1960 there was consider- able caution in predicting what would happen in the markets during 1961. A few examples are included below and it should be noted that these are representative of many broker and adviser analyses of the same period. In , as the Dow-Jones Industrial Average began its 150-point advance, one market analyst said, "a maiority of grouos and individual issues in the stock market should be avoided." The "Stock Trend Service" headlined its November 29, 1960 analysis, "OVEREXTENDED SPECULATIONS--PIVOTAL S T O C K WEAKNESSIS BEARISH." A "Special Report" of one firm dated December 29, 1960, said in part, "* * * stocks may falter before February * * * a critical period looms ahead." The analyst for another firm as of December27, 1960, saw "lower prices in the first. quarter of 1961 with a probable significant penetration of the 1960 lows." "Value Line" urged caution in , stating that stocks "will continue to be dangerously overvalued." Despite these cautions about the market in general, most brokerage firms and advisory services continued to make specific stock purchase recommendations, is These recommendations were frequently for stocks which will "run counter to the trend," or "stocks with strong characteristics." Typical was the commentin a January 3, 1961, mar- ket letter that "* * * a rampant bull market may be a long way off * * * it is the off-the-beaten-track company that appreciation- minded investors must be seeking. * * *" However, by the spring of 1961, as the averages soured, majority opinion amonginvestment analysts had shifted to a strongly "bullish" tenor. Typical were the following comments" ¯ * * the market is still in the relatively early stages of an advancing phase * * * [April 5, 1961] Mostly, stocks have advanced because either present or near-future company earnings prospects justified their rise * * * the market has not advanced un- healthily ¯ * * the evident faith of the people in the new administration is responsible in no small measure for the almost insatiable appetite of the public to own stocks. [] On the basis of the outlook as we see it--psychological, political, financial and economic--there appears the probability of a substantial further advance in stock prices o,~er the next several years--possibly to around the 950 level on the Dow-ffones Industrial Average by 1963. (March 1961) ¯ * * the market shows few of the earmarks that characterized former periods of excesses. [April 10, 1961] Another Reason Why the Dow-J’ones Industrial Average May Reach 1,000. [ffuly 31, 1961] By the end of the year the ma.iority opinion was strongly bullish for 1962 with new highs forecast, for the year ahead. "Investors Intelli- gence," an advisory service that "digested" other services, said in its January 8, 1962, issue" Consensus opinion still indicates new highs for 1962. Majority opinioa at the turn of the year suggested strong bullish trends in the first half. But a growing segment of contrary opinion sees the big 1969 bull move deferred until the second half.

See ch. III.C.4.a. REPORT OF SPECIAL STUDY OF SECURITIES MARKETS 831

Although there were a few notable exceptions, the majority of in- vestment advice contained in brokerage market letters and subscrip- tion publications continued optimistic th~o_ughout the first months of 1962. All of the eight selected stocks studied by the Special Study were strongly recommended as "buys" by an overwhelming majority of14 investment advisers.

Although subsequently overshadowed by the abrupt and unexpected 1-day break of Monday, May 28, the 5 preceding trading days were most unusual. Price declines were severe and volume was high. The Dow-JonesIndustrials declined eve~r~.~ day of the week for a net drop of 38.83 points, 4.19 percent. The SEC 300 Stock Index declined 6.8 percent. On Friday, , the Dow-Jones Industrials closed down nearly 11 points, and the tape was 32 minutes late at the close. This was one of the most severe weekly declines in market history. Every industry group, as measured by the components of the SEC Indexes, fell and most fell substantially. On Friday, , 1,334 of the 1,544 individual securities listed on the NYSEwere traded, higher degree of participation than on any previous day that year. Of those traded, 1,004 declined, 170 rose, a~d 160 were unchanged. There were 695 new 1962 lows and only 8 new highs recorded during the week. Reported NYSEvolume totaled 22,989,000 shares, several million more than in the preceding week and the most active week in 1962 up to that time. Odd-lot short sales reached a record figure for many ears. Nevertheless, the overall volume for the year to date still ~agged far behind 1961 when the market was moving upward. The magnitude of the week’s decline provided extensive comment on the future and had a clear impact upon investor psychology. Th’is was evident from the Sunday, , edition of the NewYork Times. The late Burton Crane, stock market columnist, raised the question whether the market decline of the previous week was the "* * * fore- cast of a downturn in the business cycle or ~ mer~ readjus~nent in the price of glamour stocks." He also said it raised the .question for investors of "whether to get out of the market or to dig in and hope for the best." The article further stated that professionals seemed to be advising to get out and that the best loss was the first loss. The New York Times financial editor, John G. Forrest, noted the absence of "panic selling among investors" as the one encouraging factor in the prior week’s "dismal performance." He quoted Keith Funston, president of the NewYork Stock Exchange, as saying that investors were not panicking but that there ~vas "a considerable diminution of their confidence." Edwin Posner, chairman of the board and president pro tern of the American Stock Exchange, was also reported as saying that an "adjustment" was taking place in the market and that stocks ~vere getting down to "realistic levels." One New York Stock Exchange firm in a sizable advertisement recom- mended the purchase of certain stocks, "Far Below Their Highs." An advertisement for an investment advisory service headlined its mes- sage, "Another 1929 Stock Market

See app. XIII-B, charts la through lb. 832 REPORT OF SPECIAL STUDY OF SECURITIES MARKETS In addition to the differing reports in various newspapers and mag- azines, investment advice issued over the weekend by brokerage firms and advisory services was divided. Few foresaw a break of the magni- tude that was actually to occur although some correctly anticipated a further decline. On the one hand-- Goodrecovery must be somewhere near at hand. Suggest it is time to buy stocks for sizable rally. Consider that selling has been overdone. * * * selling, largely emotional, has been overdone against the present busi- ness background * * * a turn may not be far off. Many equities have declined to zones of good support. And on the other hand--- Themarket is still grossly overvalued despite its recent decline. * * * the averages must descend further * * * some stocks will require years to attain their erst~vhile strength and price levels. On the widely disseminated Dow-Jones News Tape, commonly known as the broad tape, the followir~g commentappeared prior to the open- ing on Monday morning: * * * Should the industrial average be able to hold around the 610 level, brokers say the dramatic effect alone may prove of near-term help to the market generally--if it should not, however, these sources say that the technicians would be groping for a clue as to ~vhere a solid bottom might be found. Such was the background "opinion" before the opening on Monday morning, May 28. C. THE DAYS OF TttE BREAK The NYSEin its study of the market break detailed an aggregate picture of the events of May 28-31, which need not be duplicated here. Instead, the crucial events are summarized with the addition of a few observations and some further background material.

1. :IKONDAY,I~AY 2 8, 19 6 2 The tape began to run "late" within the first half hour of trading on the NYSE; and AT & T, often watched as a market barometer, was selling at 1091/~, ex-dividend, at 10:20 a.m., off 21/2 points from Friday’s closing. However, it was not until noon that the unusual character of the day finally became clearly evident. At that time, the Dow-Jones Industrials were reported down 10.98 points to 600.90. By 2 p.m. the Dow-Jones Industrials were r~ported on the tape at 598.06, although when the average was subsequently computed on the basis of actual transactions it had actually declined consider- ably further to 591.25, a difference of 6.81 points. At 2:08 p.m. the tape was reported 52 minutes late, one of the greatest lags in its history. However, a transaction in AT & T which took place at 2:10 p.m. did not appear on the tape as a regular report until 3:30 p.m., indicat- ing that actual tape lateness was considerably more than publicly re- orted. At the close, 3:30 p.m., the tape was reported 68 minutes late, ~ut the last trade of the day ~vas not printed out for 2 hours and 9_2 r~inutes, at 5:58 p.m. For the day the D(>w-Jones Industrials closed at 576.93, off 34.95 oints or 5.7 percent; the utilities closed at 104.35, off 8.22 points or ~.3 percent; the railroads closed at 122.35, off 6.88 points or 5.3 percent. Although in terms of points the decline was exceeded only by the 38- REPORT OF SPECIAL STUDY OF SECURITIES MARKETS 833 point drop in the Dew-Jones Industrials of October 28, 1929, in terms of percentage decline the drop had been exceeded on 23 other days since 1927. The reported volume amounted to 9,350~000 shares~ the fifth most actinic day in the history of the NYSEto that date. Of the 1~544 issues listed on the NYSE~1~212 closed lower than on Friday~ May 25. Nine hundred and thirty-seven stocks reached new lows for the year and only five~ all preferred stocks~ touched new highs. The most active stock on the NYSEwas AT & T which~ on a volume o,f 282~800shares, declined 11 points to 1005~its lowest level in 2 years. The hectic activity on the Exchange prevented even the members on the floor from keeping track of their own transactions. This con- fusion is reflected in the large number of transactions which were made to offset errors in executions. Though the total number of such transactions is not available, this situation is reflected in the fact that the aggregate volume and members’ trading data~ submitted by the Exchange to the SEC~were revised three times in the 9-month period following~ the break. The area of greatest change was in specialists transactions,15 especially sales. The volume of trading was such that there were substantial delays in the execution of public orders. Some orders were executed at prices substantially different from those which prevailed when the order was entered. In some cases instructions by customers to cancel orders previously entered were also delayed in reaching the floor until a, fter the order was executed. On ~ 1962~ the :NYSEBoard of Governors adopted a special rule governing the obligations of member firms with respect to the execution of orders during the week of May 58. In effect this rule purports to exculpate memberfirms for delays not caused by "negligence" in executing orders. The confusio.n on the floor also greatly hindered the execution of odd-lot orders. Manyorders were not "time-stamped" and because of the overwhelming volume the odd-lot orders in several issues were sorted in the smoking rooms rather than at the trading posts. Market orders o.ften were not executed at the first round-lot sale following receipt as required~ but some time later~ and on May 28 many unex- ecuted orders received earlier during the day were executed at the closing price~ plus or minus the odd-lot differential. On the American Stock Exchange the magnitude of the general price decline was slightly greater than on the NYSE.The New York Herald Tribune’s American Stock Exchange Index of 25 stocks dropped 6.3 percent, as compared with the Dow-J~nes Industrial per- centage decline on the NYSEof 5.7 percent. However~ volume on the Amextotaled only 2,980~000 shares~ a figure which had often been exceeded in 1961 and which was far from the record 7~096~300 shares traded on October 29~ 19~9. In~ the over-the-counter market May 28 was a day of confusion. The absence of accurate, current information as to prices and volumes

~ In the set of figures submitted in June 1962 the Exchange indicated that specialists purchased 7,319,530 shares for the week of May 28 and sold a total of 8,26~8,030 shares of which 1,62:8,760 shares were sol4 short. In the third revision of these data, sent in , the Exchange indicated that specialists had total purchases of 7,342,390 shares during the week and sales of 8.345.110 of which 2,047,100 shares were short sales. Specifically, the specialist in Brunswick reported making 24 transactions totaling 3,000 shares to offset errors on May 28, 29, and 31 ; the specialist, in Korvette reported 15 s,uch transactions totaling 2~900 shares; and the specialists ia U,.S. Steel reported that such transactions totaled 5,600 shares. ~ See oh. VII.C.2. 834 REPORT OF SPECIAL STUDY OF SECURITIES I~ARKETS in over-the-counter securities tended to inhibit trading. The volume of transactions was reported as significantly more than the imme- diately preceding period but less than in May1961~ a period of extreme ~ctivity. The lateness of the NYSEtape and the size of the price declines on the NYSEprompted some over-the-counter dealers to withdraw as market makers in certain securities. Others simpl.v re- duced prices in sympathy with the declines reported on the NYSE. Dealers substantially widened their price quotations and quoted "sub- ject" markets. Communication between firms was poor~ further con- tributing to a loss of liquidity and continuity.

The market opened on May 29 under the pressure of heavy waves of selling. Openings in AT & T, Brunswick, IBM, Korvette and other issues were "delayed" because of large amounts of sell orders. At 11 a.m. the Dew-Jones Industrials were reported off 11.09 points, or 1.9 percent. Nevertheless~ in the morning several major brokerage firms 17 sent wires to their branch offices suggesting that prices and yields indicated good "buys" in certain "quality" stocks. At 11:02 a.m. AT& T opened at 98~/2~ off 21/8 from Monday’s clos% on two blocks of 50,000 and 10,000 shares. At 11:40 transactions took ~qlaceuality" in issuesAT & T a.lsoat 100 reversed and at 12:23their downwardp.m. at 101~/2.direction Several in terms other of actual transactions on the floor. At 1 p.m., however, the Dew-Jones Industrials were still reported in a declining trend, down13.61 points from the open.ing. As subsequently reconstructed from actual trans- actions occurring at that time~ the average was really up 7.72 points. Because of.the late tape the gap continued througho.ut the day between the~s reported averages and the actual transactions. The rise in prices continued throughout the day as buying orders from the public began to appear with greater frequency. At 3:30 p.m. the ticker was reported ~ hours and 23 minutes late. Flash prices at that time showed AT & T at 1081A~ up 7~/s; Standard Oil of New Jersey at 503/~, up 4a/s; IBMat 378~ up 17; U.S. Steel at 53~/~ up 3; and General Motors at 50%, up 1~/~. The Dew-Jones Industrials at the close were at 603.96~ up 27.03 points, or 4.7 percent. The Dew-Jones Railroad Average closed up 2.18 points, or 1.8 percent~ and the utility average closed up 5.38 points, or 5.15 percent, its biggest gain in 1 day since November13, 1929. From its closing level on Friday~ May 25~ until 12 o~clock Tuesday, the Dew-Jones Industrials had plummeted 48.64 points to 563.24. The massive recovery of the afternoon trading recouped 40.72 points, or about 84 percent of the decline of Mondayand early Tuesday. Volume was reported at 14,750~000 shares for the day~ the second greatest volume in history~ and reater number of stocks were traded than on any previous day. (~fgthe 1,544 listed issues~ 1,399 were bought and sold. The early decline pushed 1,032 stocks to new 1962

x~ Among these firms were the following: Francis I. du Pont & Co. : Eastman Dillon, Union Securities & Co. ; tI. Hentz & Co. : E. F. Hutton & Co. ; Carl M. Loeb, Rhoades & Co. ; Merrill Lynch, Pierce, Fenner & Smith, Inc. ; Paine, .Webber, .~ackson & Curtis; Walston & Co., Inc. Garfield Drew, publisher of the advisory service, "The Drew Odd Lot Studies," sent a night letter Monday night to his wire-subscribers urging them to buy "for a substantial rally." ~ NYSE, "The Stock Market Under Stress," pp. 11-12 (1963). REPORT OF SPECIAL STUDY OF SECURITIES MARKETS 835 lows before the rise in prices began. At the close, 630 issues were higher than on Monday~637 were lower, and 132 registered no change. Twopreferred stocks hit new hig.hs for the year. American Stock Exchange prices and volumes moved generally in tandem with the NYSE. The New York Herald Tribune’s American Stock Exchange Index closed at 118.68, up 3.05 points from Monday. Volume was high at 5~330~000 shares, and 922 stocks were traded~ with 306 advances~ 480 declines and 136 unchanged. There ~vere no new highs and 672 new lows were registered. Price spreads remained wide in the over-the-counter market and the limited number of stocks traded had little depth. There was an in- creased amount of public buying, but prices did not appear to rebound to the extent they did on the NYSE. Some of the over-the-counter firms interviewed by the study said they used the stepped-up buying interest exhibited by the public "to lighten up" their positions in several securities in their inventories. As on Monday, prices were uncertain and quotations wide.

3. WEDNESDAY, lV[AY 30~ 1962 As a result of the high volume of orders to be cleared and the in- ordinate amount of bookkeeping involved, most major brokerage firms stayed open and were unusually busy over the Memorial Day holiday. The NYSEhad suggested to its members that they ignore the holiday in order to process the mountains o:f orders. Somefirms reported that their phones were also busy on Wednesdaywith a large number19 of "buy" orders.

4. THURSDAY, MAY 31~ 1962 The accumulated "buy" orders over the holiday were presumably responsible for the sharp price rise and high volume at the opening on Thursday. By 11 a.m. the Dow-Jones Industrials were up 9.03 points at 612.99~ but down from the opening high of 620.54. AT & T~ the day’s most active stock~ opened at 115, drifted to 1103~and finally closed at 1131A~up 4~ points from Tuesday’s closing, on a volume of 414,800 shares. The Dow-Jones Industrials closed stt 613.36~ up 9.40 points or 1.6 percent, thereby recovering the losses of the week. There were 1,357 issues traded with 1,071 advancing, 190 declining and 96 unchanged. The rails closed at 129.19, up 4.66 points~ or 3.7 percent. The utilities closed at 113.54~ up 3.81 points~ or 3.5 percent. Volume was 10,710,000 shares with the last trade appearing on the tape at 5:25 p.m. An unusual feature of the day was the exceptionally high volume of odd-lot purchases which totaled 1,99~4,638 shares with a market~value of $93,704,771 as against sales of only 679, 944 shares with a market value of $37,161,415. 2° This was an abrupt turn to the buying side by the odd-lot customers in contrast to their sale balances of Monday and Tuesday. On the American Exchange volume declined to 2,840,000, almost. half of Tuesday’s volume. The New York Herald Tribune’s Index

The Wall Street .lournal, May 31. 1962, p. 3. SE(~ Statistical Bulletin, July 1962, p. 18. 836 REPORT OF SPECIAL STUDY OF SECURITIES MARKETS closed at 122.25, up 3.57 points. Of the 903 issues traded, 676 a.d- vanced, 120 declined and 107 were unchanged. One new 1962 high was registered and there were 25 new lows. Over-the-counter trading on Thursday followed the same pattern as Tuesday, with prices apparently substantially unchanged.

5. REACTION OF FOREIGN :M:ARKETS By May 28, the stock markets in , Germany, Great Britain, the Netherlands, , , and had already declined further and faster over the past year than in the United States. On the other hand, the French and Austrian markets were still rising. Although foreign markets tend to go their own ways, they did react quickly thoughbriefly to. the events on the NewYork markets during the days of the break. On May 28, for example, the market closed quietly, o~le- half hour after the opening of the NYSEand before the dramatic, downward plunge was apparent. Because of the time differential the reaction to the Mondaydecline did not take place until Tuesday. Measured against the Monday 5.7 percent decline in the Dow-Jones Industrials, the London market on Tuesday dropped 6.7 percent, Frankfurt a/M () 5.3 percent, 4.6 percent; and Amsterdam3.2 percent. In similar fashion, the foreign reactib.n to the Tuesday reversal and upward trend took place on Wednesday when price recoveries virtually madeup all of ~he previous day’s losses. Thereafter, the major foreign stock markets continued the relatively independent courses which they had been following previously.

6. EPILOGUE The unusual and dramatic week closed on an anticlimax. Fri- day’s volume declined to 5,760,000 shares and individual stock prices showed small variations with almost as many advances as declines. The averages closed off slightly. Many stocks were virtually un- changed for the week. Over the next several weeks stock prices drifted lower on relatively high volume. On the Dow-Jones Industrials touched 52¢.55 for the intraday low of the year, and the next day closed at 535.76, which was the lowest closing for the year, a decline of 27 percent from the high set in December 1961. This ~vas a larger percentage decline than that of the other post-World War II bear markets: 1946, 25 per- cent in 5 months; 1949, 16 percent in 6 months; 1953, 14 percent in 9 months; 1957, 20 percent in 6 months; and 1960, 18 percent in 10 months.21 Ex. cept for 1946, each of these other declines anticipated an eco- nomic recession of varying severity, but despite its comparative sharp- ness the market break of 1962 did not anticipate or presage such an event. Instead, the economy continued its sl.ow but steady advance through 1962. By the end of the summer, however, the impact of the market break was felt directly by the securities industry. Trading volume and pub-

u~ However, the average rate of decline per month was not as severe as either the decline of 1946 or 1957. NYSE, "The Stock Market Under Stress," p. 3 (1963). REPORT OF SPECIAL STUDY OF SECURITIES :MARKETS 837 lic interest gradually subsided. By August, NYSEvolume was off 10 percent from 1961. The "small investor," if the odd-lot figures are representative, seemed clearly disinterested in the market. The vol- rime of odd-lot trading declined to an unusually low level.. Over the long term, the ratio of odd-lot trading to total volume has averaged about 9 percent. 2~ In May 1962 it was 8.9 percent, dropped to 8.1 percent in June, and by Novemberhit a low of 6.7 percent. In addi- tion, after July there was a marked excess of monthly sales over pur- chases. The sales, of mutual fund shares also de~clined in the months following the break, although the rate of redemptions did not increase as markedly. The new-issue market ~.’as particularly hard hit by the market break. The number of new commonstock offerings dropped from 273 in the first quarter .of 1962 to 211 in the second, 99 in the third, and 101 in the fourth. The first quarter of 1963 was the lowest quarter for the underwriting of new issues since 1958. The smaller companies par- ticularly abandoned previously proposed stock issues. Only 76 issues of under $2 million were filed in the third quarter, and another 76 in the fourth, as against 166 in the second quarter and 228 in the first. The general public disenchantment with the market affected the profits of the brokerage houses. By the end of the year, many sales- men had left the securities industry. 23 Smalldr over-the-counter houses ~vere the first to suffer; some quit voluntarily and a few went into bankruptcy. ~4 Even among the NYSEmember firms the pinch ~vas felt. The net income in 1962 of the largest firm, Merrill Lynch, Pierce, Penner & Smith, was about one-half of what it had been in the preceding year. Some firms merged and others embarked on cost- cutting programs; several trading and clearing departments were closed, training programs curtailed, and research staffs reduced. The New York Society of Junior Analysts reported that 50 of its 400 members had become unemployed. ~5 Subscriptions to advisory pub- lications which had been declining since early 1961 dropped off even further. The cut in margin requirements from 70 to 50 percent announced by the Federal Reserve Board on July 9 was followed by a brief upturn in the market ~verages. However, the really dramatic turnabout in the price averages began in late . By , on the anniversary of the break, the Dow-Jones Industrial Average was close ~gain to the alltime high of December1961. Just as the break of May 28, 1962, was unusual and generally un- expected, the pace of the subsequent price recovery was equally unusual. D. ACTIVITIES OF THE PARTICIPANTS This broad-brush treatment of the market break and the periods im- mediately preceding and following may be helpful in understanding the environment in which it occurred. To evaluate the factors lead- ing to the break, conditions during the break, and steps that might be

e-~ The odd-lot ratios here and throughout this chapter have been co.mputed as the ratio of odd-lot purchases and sales to the sum of these odd-lot purchases and sales, plus round- lot purchases and sales. See ch. VI.E. ~ See eh. I, table 1-18. e* Business Week, Aug. 25° 1962, p. 82. ~ New York ~imes, Aug. 11, 1962, p. 21. 838 REPORT OF SPECIAL STUDY OF SECURITIES MARKETS taken to ameliorate these conditions, it is necessary to ascertain more closely who was buying and selling stocks and, to the extent possible, the timing of these transactions relative to the price changes in par- ticular issues. As indicated above, this analysis is in terms of aggre- gate (NYSE) activity and activity in eight selected stocks. For this purpose, the participants in the market and their activi- ties before and during the break are analyzed under two main head- ings: public participants and members of the New York Stock Ex- change. The ,group of public participants, hereafter referred to as "nonmembers," is a large and heterogeneous group classified herein as (1) public individuals, (2) foreign accounts, (3) financial institutions, and (4) open-end investment companies. 26 The members of the NYSEdiscussed are (1) specialists, (2) floor traders, (3) members off-floor, and (4) odd-lot dealers.

:Nonmember activity usually dominated both sides of the transac- tions in the eight stocks selected for study (tables XIII-2a through XIII-2h), but their buying and selling varied widely among these stocks and in each stock separately showed considerable variation from day to day. In the case of IBM, member purchases exceeded those of nonmembers in 8 of the 16 nonbreak days. It is difficult to draw general conclusions concerning the activity of nonmembers relative to the broad movements of stock prices. In the first 5 months of 1960, for example, with a 54-point drop in the Dow- Jones Industrials, nonmembers, in round lots, sold 242,390,000 shares and purchased 146,367,000. By way of contrast, in the same 5 months of 1962 when the Dow-Jones Industrials dropped 111 points, the non- members bought 308,200,000 and sold 304,139,000 shares. On a weekly basis, between and , 1962, when stock prices experienced their steepest decline and reached a low fdr the year, nonmembers had sale balances in 4 out of the 8 weeks and continued to have a sale bal- ance in the 3 weeks of abruptly rising prices that followed (chart XIII-e). At least to some extent, this indeterminate relationship be- tween balances and price movements may be explained by the dis- parate nature of the nonmember groups. In the following section, therefore, the transactions of these ma.ior components are analyzed, starting with the public individuals. The analyses of these components are based exclusively upon the transac- tions entered by that segment of the respective groups covered by the Special Study. a. Public individuals (1) A ~rregate aetivlty In both listed and unlisted stocks individuals as a group were net sellers in the first quarter of 1961 ; net buyers~ by a very slight margin, during the second quarter; and net sellers once more during the third

m Tables XIII-3a through XlII-3h give a breakdown of trading by public individuals. foreign accounts, and "others" in order to provide for the full reporting of data collected from 25 broker-dealers. (See app. XIII-A.) ’The group "others," is not analyzed a separate category because of its heterogeneity. It lnclud~es, among others, nonmember broker-dealers, c(~mmercial banks, trust companies, personal holding companies, partner- ships, and nonflnancial corporations. It also includes financial institutions and open-end lnvesl:ment companies which are analyzed separately on the basis of more complete data. REPORT OF SPECIAL STUDY OF SECURITIES MARKETS 839 and fourth quarters. During the first half of 1962 individuals sub- stantially increased the volumeof their net selling, 2~ while institutions, by comparison, increased their net buying. According to the NYSE,individuals accounted for approximately 57 percent of total round-lot and odd-lot volume over the 3 days of the break--an unusually high amount. On May 28, they had a sale bal- ance of 1,572,000 shares and on May 29 a sale balance of 1,111,000 shares. A reversal of this situation occurred on May31 as individuals’ purchases exceeded sales, resulting in a purchase balance of 1,946,000 shares. 2s Odd-lot customers were particularly active during the week of the break and odd-lot buying in total and on balance was unusually high on May31, with almost three shares bought for every one sold. (2) Activity in the eight ~tock~ With respect to the selected stocks, the acti~ties of individuals were str.ikingly varied. ~9 In general, during the 16 nonbreak days--14 pmor .and ’2 subsequent to the break--individuals were the dominant nonmembergroup in respect to both purchases and sales in each of the eight stocks, with the exception of the purchase activity in General Motors and Standard Oil. In Avco, Brunswick, and Korvette they completely dominated both purchases and sales and the level of their activity was about even on both sides. In AT & T, General Motors~ IBM, and Standard Oil, however, their selling activity w~ts clearly dominant; and, conversely, their purchases were strikingly more dominant in U.S. Steel. On individual days during the break period the net buying and selling of individuals in particular stocks departed sharply from the pattern of th~ 16 nonbreak days as well as ~rom the overall pattern for the 3-d.a~ period (tables :~III-3a through XIII-3h). In gen- eral, on May28, individuals entered a majority of their sales through "market" orders and a majority of their purchases through "limit" orders: conversely, on May 31, individuals entered the majority of their purchases at market and greatl~ increased the use of limit orders when selling (table XIII-4). Finally, excluding the transactions IBM, individuals appeared to make relatively greater use of margin accounts in the stocks of lesser quality. The following are thumbnail descriptions of the activities of in- dividuals in the eight stocks. (a) American Telephone d~ Telegraph Co.--During the 16 non- break days, public individuals had consistent net sales in AT& T; oa only I day, April 27, was there a slight purchase balance. On May28, the first day of the break, individuals sold 84,300 shares, accounting for 86 percent (~f the nonmembersales, and bought 51,400 shares, or 76 percent of their purchases. In ’addition, this dominant position on the sell side was heightened by the fact that 73 percent of the round-lot (and 80 percent of the odd-lot) sell orders were "at market." The combination of a heavy sale balance and sales primarily "at market" placed great downwardpressure on the stock. Individuals continued to sell more than they bought on May ~9 and May 31, days

~ SEC, "Volume and C(~mposition of Individual Savings." ~ NYSE, ’"The Stock Market Under Stress." p. 14 (1963). :~ The transactional data inclu4ed in this ,section and the section on foreign trans- actions are based on information obtained from the 25 largest NYSE commission firms, For a complete description of these data, see app. XIII-A. 840 REPORT OF SPECIAL STUDY OF SECURITIES MARKETS on which the price of the stock rose. In AT & T, the bulk of both buying and selling activity came from cash accounts on May 28 and 29. There was a decided increase in the use of market orders to buy on the 29th and 31st with a corresponding decrease in the use of limit orders. Market orders for sales were again dominant on the 29~h and stop orders to sell dropped from 8 percent of sales on the 28th to 1.7 percent on the 29th and dropped almost to zero on the 31st. On May 31, the distribution between limit and market orders to sell was about equal. Over the 3-day period the bulk of individuals’ total actL~ity and net sales emanated from cash accounts. (b) Avco Corp.--For the 16 nonbreak days the number of shares purchased by individuals averaged 82 percent of nonmember pur- chases and 87 percent of their sales. On ~ay 28, individuals sold more than they bought, accounting for 85 and 81 percent of nonmem- ber purchases and sales respectively. Eighty-one percent of their round-lot sales were market orders, and 8.3 percent were stop-loss orders; on the 29th and 31st individuals were net buyers with both purch.ases and sales.preponderantly on market orders. In Avco round- lot sales from margin accounts exceeded those from cash accounts over each of the 3 days of the break. (c) Brunswiclc Corp.--For the 16days, individual activity averaged about 90 percent of both nonmember round-lot purchases and round- lot sales. On the 28th~ share purchases and sales of individuals were about equal; in percentages, their purchases increased to 93 percent of reported nonmember volume and their sales rose to 96 percent of reported sales. Seventy-five percent of their round-lot sales were market orders and 2.4 percent stop orders; 67 percent of their pur- chases were limit orders and only 33 percent market orders. Purchases were substantially in excess of sales on May 29, and raarket orders to buy jumped ~rom 33 to 73 percent of their total. Sales exceeded pur- chases only on May 31 and market orders to sell dropped to 54 percent of their total from 75 and 79 percent market orders on the preceding 2 days. (d) General Motors Corp.~Among nonmember transactions during the 16 days, individual purchases, 40 percent, were the lowest of the eight issues while their sales, 81 percent, were relatively high. On May 28 and 29 their proportion of purchases increased somewhat and sales remained at a high level. Thus, they were heavy net sellers on May 28 and 29, principally on market orders. On May 31, individuals turned to the buying side, again principally on market orders. The bulk of both buying and selling transactions came from cash accounts. (e) International Business Machines Corp.--Although the buying and selling pattern of individuals in IBM varied among the 16 days, on the average the number of shares bought over this whole period represented 56 percent of nonmember transactions, while sales con- stituted 74 percent of the total. On the 3 break days their activity as purchasers was below the average of the nonbreak days and on May 28 and May 29 their sales activitv was above that average. On these days of the break period, indivi~duals were rather large net sellers~ employing principally market orders except for their round-lot sales on May 31, the majority of which were executed by means of limit orders. Other than May 31~ the majority of sales came from cash REPORT OF SPECIAL STUDY OF SECURITIES ~VIARKETS 841 accounts, while the majority of round-lot purchases on all 3 days came from margin accounts. On the 28th, individuals had purchases of 1,600 shares and sales of 5,500 shares in cash accounts while in margin accounts their purchases amountedto 5,300 shares and their sales 5,100. (f) K.J. Korvette, Inc.--On the 16 days, individuals dominated the reported nonmembertransactions on both the purchase and sale side, accounting for 83 percent of nonmemberpurchases and 88 percent of nomnembersales. On May 28 they accounted for 68 percent of pur- chases and 86 percent of sales. They were net buyers on May 28 and 29 and net sellers on the 31st. On May 28 the majority of purchases were made on limit orders while most sales were at market. On May 29 and 31, the majority of buy and sell orders were at market.. In Kor- vette, more than in any of the other stocks studied, the bulk of trans- actions during the break came from margin accounts. (g) Standard Oil Co. (New Jersey ) .--During 4= of the 5 days covered in 1961, individuals were net buyers of this stock, but they were net buyers in only 1 of the 11 days covered in 1962. Thus, on the average of all 16 days, they were net sellers. The average number of shares bought for the 16 days was 44 percent of all reported nonmember purchases while the average number of shares sold was 69 percent. On May’28 their proportion of purchases and sales was close to these aver- ages. During each of the 3 days of the break individuals were heavy net sellers and on the 28th 63 percent of their sales were market orders and 16 percent were stop-loss orders. Stop-loss orders were much more important in Standard Oil (N.J.) than in any of the other stocks studied. On all the days of the break, their transactions came principally from cash accounts. (h) United States Steel Corp.--This stock was the only one of the 8 studied in which individuals bought consistently on balance during the 16 days. The average volume of their purchases for these days was 79 percent of nonmemberpurchases while the average volume of their sales was only 55 percent of nonmembersales. On May 28 their pur- chases were relatively belo~v the average while their sales ~vere consid- erably above. Individuals bought slightly on balance on May28, sold rather heavily on balance on the 29th, and were large net buyers on the 31st. On ~8 the bulk of round-lot purchases were executed by limit orders but the majority of sales on May ’28 and of all trans- actions on the other days were market orders. The bulk of trans- actions on all 3 days was from cash accounts. b. Foreign participation ao (’1) Aggregate acti~ity Foreig-n participation in the U.S. securities markets has steadily in- creased in recent years, although there is considerable variation be- tween countries and in their aggregate net activity (chart XIII-f). The tim!ng and nature of ,this activity, in addition to its direct impact upon prlce levels and the supply of securities, has further significance because* of its influence on the U.S. balance-of-payments position?

a0 The term "foreigner" as used here eo.vers all individuals and institutions domiciled outside the United States. its territories, and possessions as weU as the official institutions of foreign countries, wherever such institutions may be located, and international organ- iz,ations. It is significant that a security transaction by a U.S. citizen or institution, whether here or abroad, entered through a foreign institution is also classified as a "foreign order" according to the country through which the order is entered. a~ The Special Study did not analyze the balance-of-payments aspect of foreign trans- actions in U.S. securities. 842 REPORT OF SPECIAL STUDY OF SECURITIES MARKETS

Chart XIII-f CUMULATIVEPURCHASE AND SALE BALANCES OF DOMESTICSTOCKS BY FOREIGNERS Dollars Mdhons Dollars Mdhons 1000 1000 (In total ~’~d by major countries or areas cumulative, by months, beginning January 1958)

9OO - 900

800 - 800

- 700 GRAND

- 600

500

4OO 400

300 300

200 200

LATIN AMERICA

100 100

-100 -100 1958 1959 1960 1961 196Z

SOURCE:Treasury Bullet,n, U S Treasury Dept OS.ii4? REPORT OF SPECIAL STUDY OF SECURITIES ~.VIARKETS 843 During 1950, foreigners bought $666.9 million of U.S. stocks and sold $g~4.0 n~.illion of these securities, or a total of $1,330.9 million (table XIIL-7). 3-~ By_.Jvl(~gl, s~.ock volume emanating from foreigm sources had increased more than t~ourfold to $5,811.9 million. This record amount was the result of particularly heavy purchases during the active markets in the spring of the year. During’the first half of 1969~, as prices dcc]ined, foreigners continued to buy on balance, but in June 196’2,~a for the first time in 11 months, they sold more than they bough¢ with net sales amounting to $65.1 million, contrasted with net purchases of $5~:.0 million in January 196g. As for fl:e 3 days of the break, the NYSEstudy found that for- eigners aecolmfed for 5.5 percel~t of total public volume, the highest propor~io~ ever recorded in any of the Exchange’s public transaction studies. Overall, foreigners sold more than they purchased; they had a~ purchase balance on May 28 followed by 2 days of heavy selling? (2) A.c~ivi~ in ~h,e eiyh~ stoelcs Over the 16 days, foreigners had some purchases and sales on most of the days in all of the eight stocks but the volume of their activity and the distribmion of this activity among these stocks and between purchases and sales varied greatly. Ou the average the volume of these transactions was small comparedwith that o.f individuals, insti- tutions, or membersbut, nevertheless, on particular days, foreign tivity was quite importtmt. For example, on at least one of the days s~udied, foreigners had purchases or sales that amoun’ted to as much as 19 percent of total reported nonmemberpurchases or sales in each of these stocks. Furthermore, in particular stocks, foreigners were, at times, the dominant participant. For example, in the 5 sampled days of Novemberlg’.~t, the. volume of their purchases of IBM, although never reaching 1,000 shares per day, ranged bet~ een 37 and 71 percent of total reported no!xm ember purchases. During the 3 days of the market break, foreign activity tended to accelerate, in some cases sharply. By way of illustration, on May 28 foreign purchases as a percent of total nonmemberpurchases amounted to 11 percent in AT & T, 1t percent in Aveo, 9~6 percent in IBlY[ (34 percent on May 29), and 17 percent ’in U.S. Steel. In all of these the average percentage of their prior purchases was considerably lower. Similarly, on ~.{ay 31, :foreign sales amountedto 26 percent of the nonmembersales of General Motors, compared with 8 percent on the nonbreak days; 50 percent of IBM, compared with 9 percent; and 23 percent of -U.S. Steel, also compared with 9 percent. In Bruns- wick, on the other hand, foreigners had a relatively small volume of transactions on each of the 3 days of the break. From additional data supplied to the study, it was found that there was particuiarly heavy consistent selling activity by foreigners in General Motors in late April and May of 1969, most of it coming from the United Kingdom. For example, five large U.S. commercial banks reported ~ore]gn sales of General Motors as follows"

~’-"The total value of purchases and sales by foreigners compiled by the U.S. Treasury Department is no~ necessarily complete since foreigu institutions with transactions ~ver- nging less than $100,000 per month in the preceding 6 months need not report. ,~a The June figures inclu~le transactions consummated in late May. ¯~ N¥SE, "The Stock Market Under Stress," p. 8 (1963},

96-746--63--1~t.. 844 REPORT OF SPECIAL STUDY OF SECURITIES MARKETS

TABLE XIII-a.--Sales of General Motors Corp. emanating ~rom the United Kingdom, selected days, 1962 Sales (in ~hare~) (in shares) 1962---Apr. 27 ...... 5, 100 1962-- 3,000 30 ...... 12, 100 28 ...... 2,000 May :1 8, 800 31 ...... 6,000 11 ...... 4, 900 Since these transactions were reported by broker-dealers as those of a domestic commercial bank, they were not classified as "foreign activity." Thus the available data from broker-dealers on foreign trading are inadequate for a proper and regular evaluation and under- standing of the role of this important element in the securities markets. c. Institutiona~ Tartieipation 3~ (1) Aggregate activity Institutional activity has been an increasingly important factor in the markets for commonstocks, both in terms of absolute and relative activity. On the basis of materials analyzed in chapter VIII and this chapter it is apparent that in the aggregate institutions tend to accum- ulate stock on a net basis. In every month from September 1961 through June 1962 the 162 institutions that were studied had a pur- chase balance on the NYSE(table XIII-8). Moreover, as prices generally began their downward slide, institutional purchases rose noticeably while sales, beginning in March 1962, remained relatively unchanged. As a result, net purchases ranged from a low of 0.4: per- cent of NYSEvolume in December 1961 to a high of 2.6 percent in June 1962. Weekly purchases by institutions generally increased from the end of March 1962 and reached a peak of 1,079,5~8 shares in the 4-day week ending (table XIII-9). Sales of 171,724 shares in that week were the lowest for the preceding 2 months. Thus, resulting net purchases of 907,824 shares represented the largest purchase balance by the institutions in any week for the period studied. Throughout June, the institutions continued their pattern of buying heavily on balance. Even on an aggregate basis, however, it is apparent that the different classes of institutions did not act alike and in some cases pursued dif- ferent courses of action. For example, from September 1961 through March 1962, college endowments consistently sold stocks on balance. In December 1961, nonlife-insurance companies had an unusually heavy sale balance compared with an unusually heavy purchase bal- ance by the pension funds. In June 1962, when the institutions as a group attained peak net buying, the closed-end investment companies were sellers on balance, principally because two important, companies were selling heavily. (2) Actisity in the eight stocks Whenanalyzed on a stock-to-stock basis, the activity of institutions as a group also varied (tables XIII-10a through XIII-10e). With

~ The institutions discussed here are those which responded to questionnaires IN-l, IN-2, and IN-3 and include common trust funds (bank administered), foundation,s, closed- end investment companies, college endowment funds, pension funds, life insurance com- panies, and nonlife-insurance companies. Open-end investment companies are analyzed sepa rat.ely. REPORT OF SPECIAL STUDY OF SECURITIES MARKETS ~45 respect to General Motors, AT & T, and Standard Oil the institutions studied were for the most part predominantly net buyers prior to, during, and subsequent to the break days covered and did their maxi- mumnet buying on May 29. During particular periods they absorbed, on balance, 8 to 11 percent of total reported NYSEvolume. The institutions’ highest relative activity was in IBm{, where they were heavy net sellers from September 1961 through January 1962; they were also net sellers of IBM on the selected day in and in May 1962. In some instances, their sale balances were unusually large, reaching as high as 24 percent of reported NYSE volume in December 1961. Although the institutions had small sale balances in IBM on and 14, they had comparatively high net purchases on each of the 3 days of the break. Activity in U.S. Steel followed still a different pattern. From September through December 1961, the institutions were heavy net buyers, but shifted to the selling side in January 1962. From that time through most of June, they had net sales which went as high as 11 percent of reported NYSEvolume for the week ending . The institutions had a minor purchase ba]ance on May 128, a some- what larger one on the 29th, and a still larger balance but this time on the sale side, on the 31st. In Avco, Brunswick, and Korvette, the institutions had relatively little activity. el. Open-enel investment co~panies a~ (1) Aggregate activity The purchases of the open-end investment companies did not exceed sales by much of a margin h’om September 1961 until March 1962. In September 1961, the open-end companies reported a small sale bal- ance (table XIII-11). From March through May 1962, the net buy- ing of open-end companies began to accelerate and in this respect. their activity was similar to that of the other institutions previously discussed. They did their heaviest buying in the week ending MayItS, just~ before the break? During the week of the break, while other institutions increased their purchases and reduced sales, the funds curtailed their purchases slightly compared to the previous week, and increased their sales by a substantial amount. The net purchases o.f the open-end companies represented a smaller proportion of reported NYSEvolume than did the net purchases of the other institutions, despite the fact that ~he volume of activity of the open-end compames was substantially higher. (~) Activity in the eight stocks The open-end funds participated in each of the eight stocks (but were particularly active in the five which had been fund "favorites" for some time; namely, AT & T, General Motors, IBM, Standard Oil and U.S. Steel (tables XIII--12a through XIII-12h). The fol- lowing table shows the changing rank of these 5 stocks according to a survey of the purchases of 350 investment companies.

~ Data discussed he~ are based on respon,ses by 55 open-end investment companio~ to questionnaires IN-l, IN-2, and IN-3. ¯ ~7 Naturally, not every fund bought stock on balance and certain individual funds were strongly "bearish," selling stock and building liquid balances. ~46 REPORT OF SPECIAL STUDY OF SECURITIES MARKETS

TABLE XIII-b.--Ranlcings of 5 stocks held by funds, selected periods, 196I-6~

Rank by dollar value Statistics on June 30, 1962

1961 1962 Shares Percent Dollar Number held outstand- valt:m of funds (thou- ing stock June 30 Dec. 31 I Mar. 31 June 30 (millions) holding sands) held by funds

AT&T...... 3 241 132 2,320 ¯ 98 General Motors 9 240 151 4, 968 1.75 IBM ...... 1 1 480 153 1, 4t4 5.13 Standard Oil (N.J.) ...... 7 270 143 5, 2.49 U.S. Steel ...... 5 10 75 78 1,693 3.13

Source: Vickers Associates, Inc., "Guide to Investment CompanyPortfolios." As is evident from the table, the most dramatic changes in rank were the decline of U.S. Steel from 5th to 34th position in 1 year and the advance of Standard Oil from 7th to 8rd. These changes are cor- roborated by the transactions in these stocks by the 55 funds studied. In the period covered, these funds made heavy purdmses of Standard Oil and unusually heavy sales of U.S. Steel. (a) American Telephone & Telegraph Co.--AT & T was strongly bought on bala.nce by the funds, particularly in April, May, and June. On the days of the break the reporting funds bought AT & T most, heavily on Monday,May ~8, as the price drGppedfrom 1091/~ to 1005/s ; they acquired 1~000 shares (4£ percent) and sold none2s On Tues- day, May ~9, when the overall volume was considerably larger and the price rose from 98~/,~ to 108~/~, the reporting funds bought less--8,~00 shares (~.4 percent). One large firm, which was not among those in- eluded in the sample study, purchased 10,000 shares of AT & T at 100 shortly before noon on May~9. According to the testimony of one specia]ist on the floor that day, word of this transaction "spread like wildfire." On Thursday the 55 funds sold 500 shares (0.1 percent) and bought none. (b) General Motors Corp.--General Motors was accumulated on balance by the funds from September 1961 through March 1962 with the exception of December. By the first 3 weeks i~Ma.y, however, they were selling General Motors on balance, reaching a net sale level of 4"2.,700 shares (14.9 percent) in the week ending May1.8. In the week ending May 25, the funds purchased 4,700 shares on balance (2.1 per- cent). On May~8 the funds sold 6,500 shares on balance (6.5 percent). The nex~ day, as the price rose 2~/s points, they bought :~.~700 shares (1.9 percent) and sold none. Trading volulne in Gm)er,%] Motors heavier o.n that day than on any other day studied. On May 31, they sold 16,800 shares on balance (13 percent), and the price remained un- changed. During the rest of June 196’2, the funds generally sold Gen- eral Motorson balance. (c) International. Business Mac/zi~es Corp.--From September 1961, the funds generall Ysold IBMon balance- u~tfi._-" M~~rch 1962, and bou~,ht~ thereafter as the price declined. The funds bought 5,000 shares on May 28 (8.7 percent) as the price dropped 371/~ points. On May 29, as the

.~s In this discussion of fun4 activity reference will often be m:~de both to ~bsolute activ- ity, expressed in sh, ares, and to relative activity, expressed as a percent. Unless other- wise indicated, the relative activity, figures are expressed as a percent of total reported NYSE volnme. .Thes.e percentages will be shoxvn in parentheses. Chart X[ll-e WEEKLYMEMBER AND NONMEMBER GROUPS’ PURCHASEAND SALE BALANCES FROM STOCK TRANSACTIONS ONTHE NEWYORK STOCK EXCHANGE

S.E.C, INDEX OF STOCK PRICES (1957-1959 = 100)

I00 I00

SHS.THOUS&ND$ SPECIALISTS’ ROUND. LOT BALANCES

MEMBERS’ ON* FLOOR ROUND-LOT BALANCES

MEMBERS’ OFF. FLOOR ROUND-LOT BALANCES

I .I, J I I

-20

-400

NONMEMBERS’ ROUND-LOT BALANCES +400 ....

-40 -400

ODD-LOT CUSTOMERS’ BALANCES +400

d F M A M d J A 5 0 N D 1961 1962 REPORT OF SPECIAL STUDY OF SECURITIES MARKETS 847 price rose 22 points, they bought 5,100 shares (9.9 percent). They sold on balance 600 shares on May31 (1.1 percent) as the price rose 91~ points. (d) Sta~,clarct Oil, Co. (New Jersey).~The funds bought Standard O~l in increasing amounts during 1962, especially in the latter weeks of April and May~and no sales were reported by any of th~ 55 funds in those 2 months. On individual days, the ~und purchases of Stand- ~ rd Oil assumed a~ unusually large percentage of total NYSEvotume~ particularly ~n the week before the break. This stock was the only one of th~ selected eight bought consistently~ though in declining amounts, on each of the days of the break and there were no sales of the issue by ~myof the reporting funds. (e) United States Steel 5~orp.~U.S. Steel was sold on balance be- ginning in ()ctober and the last reported purchase by the reporting funds was in December 1961. In that month they sold 54~500 shares on balance (13.6 percent). In March~ the month before the steel- price controversy~ there were no purchases by the funds and they sold 3~,600 shares (17.9 percent). On the days s~udied, the fm~d sales reached the highest percentage o~ NYSE~olume on November 17, 1961~ when sales amounted to 9~900 (28.4 percent). The stock closed that day at 7~, off ~s from_ the previous close. By the end of Novem- her the stock was selling at 77~ and clo~ed for the year at 78~. De- sp.it~ the sales by the funds~ U.S. Steel~ as detailed ]n appendix XIII-B~ was strongly recommendedas a "buy" by most of the advisory services in 1961 and again in the first half of 1962. The funds sold U.S. Steel heavily on all 3 d~ys of the brea.k and purchased none. On May 28 they sold 9~700 shares~ (11.0 percent) 12~000 shares (9.3 percent) on May 29; and 1~600 shares (13.5 per- cent) on May 31. (f) The ~emai~ing three stoe~s.~The funds were generally in- ~ctive in Avco but their purchases or sales on some days constituted a significant ~maountof trading in the stock. For example, the 14,600 shares purchased on November 17~ 1961, represented 37.7 percent of trading volume ~hat day and fund purchases of 139,600 shares during were 27.3 perceng of volume for that month. Bruv_swick was sold consistently by the funds in the period before the break with no purchases reported after January 1962. Their largest sales were in the week ending April 20, when 24,000 shares (14 percent) were sold. During the break the funds sold 10,000 shares (8 percent) of Brunswick on M~ty 31 and bought none. Korvette was generally boug]~ b5, the funds, especially in the first 3 months of 1962 and ~gain in June. Fm~d purchases in Korvette on M~y28 were the highest propo~ion of ~ny of the eight stocks bought on that d~y~ amounting to 10~000 shares (16 percent)~ ~vhilo on M~y 29, 5,900 shares (5.8 percent) were bought; no sales occurred on either of these 2 days. On ! there was a s~le balance of 1~000 shares (1.2 percent) ~ a day whenthe stock closed up 1~ points. a. Me~nbersas a group In the aggregate, NYSEmembers’ transactions averaged about 20 percent of total volume each week between September 1961 and Jun~ 84S REPORT OF SPECIAL STUDY OF SECURITIES MARKETS 1962. However, the relatiYe importance of members~ trading in i~- dividual stocks varied widely from this average. For example~ in one of the eight stocks studied7 AT & T~ members accounted for 56 percent of volume on ~/ovember i5, 19617 but for only 18 percent in General Motors on the same day. Similarly 7 the relative amount of trading in the same stock differed markedly from day to day. In Avco7 for example, members were responsible for 38 percent of the trading on November 3~ 19617 but on the following trading day7 No- vember 6, they accounted ~or only 16 percent of total volume. More- over7 members may be particularly active on one side of the market, as they were in IBM when~ on November 37 1961, their purchases represented 59 percent of volume while sales were only 9 percent. In order to study ~he impact of members’ intraday trading on the prices o~ the eight stocks~ detailed charts were constructed for each stock showing the number of shares traded and the price of every sale on May 28 (charts XIII-1 through XIII-8). The tape time is shown in the charts as well as the actual time of execution for those transac- tions for which ]t could be ascertained. Individual ~ransactions of the various membergroups are identified on the charts. ~ The exact sequence of trades and the reported times ~md placing, o~ members’ purchases and sales could only be approximations, owing to limita- tions~° of .the data. The member group is composed of several clearly distinguishable sub~oToups, each of which operates under certain Exchange restric- tions and is influenced by varying responsibilities. As a result the relative importance of trading by different types of members may vary considerably overall and in individual stocks (table XIII-13). For this reason the aggregated activity of all members has limited meaning and in order to analyze the effect of the various memberson the market7 each subgroup is analyzed separately (tables XIII-14a through XIII-14h). 5. Specialists The functions o,f the specialist and his activities before and during the market break were discussed in chapter VI. This section includes some brief references to that m~terial as well as u further discussion of the intraday activities of the specialists in the eight stocks on May 28 and I~9. (1) Aggregate activity During the early part of 1962, as stock prices receded ~rom their highs reached in December1961, specialists as a group were more fre- quently net sellers rather than net purchasers each week (table XIII- 15). From the end of Decemberto 1V[~y 25, they reduced their inven- tories about one-third (chart VI-c). The extent to which this net selling contributed to the deterioration of the market prior to the actual break cannot be determined. It would depend upon the timing of the sales and the nature of the trading in the issues which they sold. On Monday~May 28, 1962~ when the Dow-Jones Industrial Average declined 34.95 points~ specialists bought 1,6~:9~800 shares and sold 1,443,400 shares, resulting in a purchase balance of 206~400 shares

~ Transaction times for nonmember groups were unavaila~ble. ~he limitations of the data and the sources (~ the sequence and timing of sales are discussed in app. XIII-A. REPORT OF SPECIAL STUDY OF SECURITIES MARKETS 849

(taMe XIII-16). The purchase bMance represented 6.7 percent their total purchases and sales and 2.1 percent of all round-lot volume on the Exchange that day. On May29, specialists sold 176,200 shares on balance as the Do.w-Jones Average gained 27.03 points. Again, this balance represented a small percentage of specialists’ trading and of total volume. In contrast to the relatively small balances of the previous trading days, on May31 specialists sold a balance of 931,800 shares--22.2 percent of their total trading and 8.5 percent of all NYSEsales. On this day the Dew-Jones Average showed a further gain of 9.40 points. As a first approach to studying intraday activity, an hourly analysis of specialists’ tradilLg on May 28 was made for the 30 stocks in the Dew Jones Industrial Average. It was found that as the day pro- gressed and the decline continued, the net purchases of specialists in these 30 stocks Vendedto decrease. Toward the latter part of the day balances~ shifted to the sale side. (2) Activity in the eight stocks Specialists were the most active membergroup in ~t]l but AT & T and IBM: during the 3 days o.~ the break and in the other 16 days as well. ( a) American Telephone & Telegraph Co.--On May 28, the special- ist purchased 3~900 shares at the opening, which was down several points ~rom the previous close. During the day he acquired 21,400 shares and sold 20~500 shares which represented 7.6 and 7.2 percent, respectively, of total round-lot purchases and sales. His activity on the purchase side was below his usual activity in this stock as measured by his average participation rate in the 16 other days, excluding the break period. On May29, the specialist began the day long 1,500 shares. He again purchased heavily--14,600 shares--as the s~ock opened down A~ter the opening until 11:30 the market in AT & T remained rela- tively stable, ran ~gmgbetween a]~proximately 98 and 99, and the special- ist purchased 2~o00 shares on ~alance. At about noon the price of AT & T and the averages moved up and thereafter the specialist sold 3,800 shares on balance as the price o~ AT & T climbed to 108~/2 at the close. For the day as a whole, he had a purchase balance o.f 10,000 shares, so that he closed the day with a considerably larger position than he generally assumed in the stock. This was in marked contrast to trading on the previous day whenthe specialist added only 900 shares to his inventory. (b) Avco Corp.--There are three specialist units in this stock and they averaged about 16 percent of total purchases and sales per day in the 16 days studied, excluding the market break period. On May 28, their relative importance as purchasers ~as close to their average, but they were responsible ~or only 6.7 percent of all shares sold. Although they had an 8,000-share purchase balance for the day, over 85 percent of their sales occurred in the afternoon when the major part of the price decline occurred. The stock opened at 21~/s, off 1/~ from Friday’s close, on an opening sale o~ 4,000 shares of which the specialists ac- quired 1,500 shares net. During the first two ’and a ha]~ hours of trading, the specialists bought 4,200 shares and sold 800 shares for

See ch. VI.D.6.e. $50 REPORT OF SPECIAL STUDY OF SECURITIES MARKETS a purchase balance of 3,400 shares. In a few instances, however, after 12:’30 4~ certain sales by the specialists may have contributed to the pressure on the stock. One instance was the sale of 900 shares at 12:47 at 205~, the lowest price reached until then on that day; in the minute period following this sale the specialists purchased 900 shares. Another instance was 2 sales of 400 and 600 shares each on zero-minus ticks 4~ at about 2:50, whenthe stock was experiencing a short but rapid decline. That downturn stopped at about 2:55 with the specialists’ purchases of 2,000 shares. At the opening on May 29 there was a large acct|mu]ation of sell orders resulting in an opening sale of 35,000 shares, of which the specialists bought 8,400 shares, opening the stocl~ at 17~/~, off 11/s. Specialists sold 11,800 shares on balance for the day with the bulk of their inventory reduction occurring after the ma.~’ket began its up- turn at about 12:30. (c) BrunswicTc Uorp.--The specialist was particularly active in this stock on May 28. Whereas on the average of the 16 days exclud- ing the market break period he bought and sold 15 percent of all the shares traded, on this day he was responsible for 27 percent of all purchases and 23 percent of all sales. The specialist in this stock was more active than were those in any of the other stocks analyzed. On May 29 the specialist continued his hea~y par~icip~tio~ in the trading of the stock. Of particular interest was his large number of short sales~" and several substantial covering purchases effected in th~ afternoon following the peak of the recovery. (d) General Motors Uorp.--There are two specialist units in this stock. Out of total volume of 99,500 shares on ~ they bought 9,200 shares and sold 2,100 shares. Their rate of participation as purchasers was about tile same as o~ the nonbreak days. One of the specialists purchased only 400 shares al~d sold 100 shares for the entire day. The stock opened at 50% and traded in the 50’s until late in the afternoon. From about 2:50 until aro~.~d 3:25, the stock dropped rapidly to 48~. During this decline, specialists bough~ 2,400 of their total purchases of 9~200 shares and sold 1,000 of their total sales of 2,100 shares. On May 29, specialists bought heavily ~t the opening of 46, down 27,/s points. Their purchases and sales for the day were about balanced and ~vere largely concentrated in the period prior to 1:00 p.m. (e) International ff, usiness Mac]dues Corp.--Of the eight stocks studied the specialist in this stock was the least active during the 16 days and on May 28 he did not increase his participaLion. On this day, he accounted for 6.8 percent of tt~.e purchases and 7.3 percent of the sales. At no time during the day did the specialist intervene in sufficient volume to slow tile rapid deterioration of the :5~arket in IBM. Moreover, many of his sales were made during periods of sharp declines and frequently more than offset any strength he might have contributed to the market by his purchases.

~ In the textual discussion of charts XlII-1 through XIII-8, transactions rather than tape times are referred to unless otherwise no~ed. ~a A "minus tick" transaction is one which is effected below the price of the 0receding transaction. A "zero minus tick" transaction is effected at the same price as the preced- ing one ~vhere that price is below the last preceding different price. "Plus tick" and "zero plus tick" transactions are those effected at a price above the preceding price and above the last different preceding price, respectively. See ch. VI.D for a discussion of how "tick test" scores are used by the NYSEto evaluate specialists’ stabilization performance. ~ See ch. ¥I.tt for a discussion of .short selling by specialists and other members in the eight stocks. REPORT OF SPECIAL STUDY OF SECURITIES MARKETS 851

On May 9‘9, the stock did not open until about 1:45 p.m. By this time, the market for many stocks had turned strongly up,yard ~nd at this point the specialist increased his long position. He purchased 3,000 of the 30,000 shares sold at the opening, and the stock opened at 3651~, up 4%points from the previous close. (f) E. J. Korsette, In.c.--Compared with his amount of participa- tion ia the 16 nonbreak days studied, the specialist’s relative activity on the purchase side was below average on May 28, whereas his sales were at about the same level. For the day as a whole, he purchased 8,800 shares and sold 12,000 shares for a sale balance of 3,200 out of 63,500 shares traded. From the opening until about 12 he engaged lnostly in small off- setting sales and purchases, but then he began to sell on balance. His selling and virtual cessation o f buying- between about noon and I p.m. was immediately followed by falling prices. As the price continued to drop, the specialist in turn bought, sold, and bought again until about 2 p.m. Thereafter, he made negligible purchases until 3 o’clock as the price plummetedto its low for the day. The specialist was exceptionally active on May29, effecting 22 per- cent of all the purchases and 24 percent of the sales. He effected large sales from 12 :~5 until 1:00 and several short sales between 1:30 and 2:30 as the stock leveled off after a rise and then declined slightly. (g) Standard Oil Uo. (New Jersey) .--There are two specialist units registered in this stock and together they were responsible for 16 per- cent of the purchases and 11 percent of the sales on May9,8 compared with 9 and 8 percent, respectively, on the average of the 16 nonbreak days. On May 28, Standard Oil experienced three separate periods of sharp decline: between 1’2:30 and 1’2:50 when the price dropped from 491Ato 48 ; between 1 :~5 and 2:00 ~vhen the price dropped from 481/~ to 47; and from 9‘ :50 to 3:25 when the price dropped from ~Ta/~ to 453/s. In each of these periods, the specialists’ purchases toward the end of the declines appeared helpful in stemming the downturns, at least temporarily. After a lower opening on May ’29, the stock movedsteadily upward. Specialists sold on balance for the day. Forty percent of their s~les were short sales and were effected in the latter half of the day. (h) United States Steel Uorp.--Compared with the rest of the mar- ket, U.S. Steel had only a minor price decline on May 9‘8. However, the stock had been under unusual selling pressure since the steel price controversy in April and had already fallen over 16 points by May9‘5. The outstanding feature of trading by the two specialist units in the stock on May 28 was the high level of their activity marked by a large volume of short selling. At 11:42, specialists sold 4,500 shares in 1 block, of which 3,300 shares were short, and continued selling short and covering a few minutes later as the price declined gradu- ally from 51aAto 50%at the close. c. Floor traders Floor traders are members of the Stock Exchange who are physi- cally present on the floor of the Exchange and who buy and sell for their own account. Only 28 members are classified as "floor traders" by the Exchange¢s but any member, other than a specialist

See oh. I, table I-4. 852 REPORT OF SPECIAL STUDY OF SECURITIES MARKETS or odd-lot dealer, who trades for his own account on the floor of the Exchangeis engaged in floor trading. Floor traders’ activity constitutes a relatively small percent of total act.ivity, but it can be, and often is, a highly significant factor in the price movement of particular stocks. Various studies by the Com- mission show that floor traders generally concentrate in a relatively small number of stocks which are both active .and tend to experience sharp price fluctuatio.ns. 46 Moreover, floor traders trade more often with rather than against the price trend, thereby adding further to the already existent price movement. (1) Aggregate activity Floor traders, like the specialists, sold on balance from the end of 1961 until the break; their net sales amountedto about 500,000 shares during the period. Their weekly activity fluctuated between 1.5 and 2.8 percent of total activity on the Exchange (table XIII-17). For the week ended Jun~ 1, 1962, they had the largest share volume of any week in the year up to that time, 1,212,500 shares, but because of the high overall volume they accounted for only 1.4 percent of all purchases and sales--their lowest ratio in the 5 months. Short selling by members and nonmembers increased in April and May and the floor trader was no exception. 47 As the market declined, floor traders’ short sales increased both in amount of shares and and relative to their total sales (t~ble XIII-18). On May 28, floor traders in the aggregate had a sale ’balance of 19,500 shares resulting from purchases of 153,700 shares and sales of 173,200 shares, of which 10,200 were short sales. Their net sales amounted to 0.2 percent of all sales that day. By comparison, mere- bers off-floor had a purchase balance of 144,500 shares, and specialists had one of 206,400 shares. (2) Activity in the eight stocks The table below indicates the relative importance of the floor trad- ers’~s activities in each of the eight stocks on May28, 1962. TABLnXIII--c.~Ratio of floor traders’ purchases and sales to total volume, May 28, 1962 [Percent]

Stock Purchases Sales Stock Purchases Sales

&T& T ...... 15. 5 14. 2 IBM...... 17.0 19.1 g.vco ...... 1.4 5.3 Korvette ...... 3.5 2.~ Brunswick ...... 6. 5 2.9 Standard Oil (N.J.) ...... 7.2 5teneral Motors ...... 6 1.9 U.S. Steel ...... 8 3.

As shown, floor traders accounted for a small portion of both pur- chases and sales of the commonstocks of U.S. Steel, General Motors, and Korvette. With the exception of sales of 500 shares in Korvette at about 1:40 when the stock was declining sharply, floor traders’ transactions in these issues were scattered throughout the day and did not appear to occur at critical times. Floor traders accounted for a somewhat higher percentage of trading in Standard Oil (N.J.)~

See ch. VI.F. See oh. VI.H., chart VI-e. Also see tables XIII-13 and XIII-14a through XIII-14h. REPORT OF SPECIAL STUDY OF SECURITIES 1V~ARKETS 853

Brunswick, and Avco and a substantial portion of the volume AT & T and IBM. Highlights of floor trading in these 5 stocks ~tre discussed belo~v. (a) An~eric,~ Telep.hone & Telegraph Co.--On the average, dur- ing the 3 days of the break, floor traders were the m.ost active member group in the stock, their participation surpassing that of the specialist. On May28, floor traders bought, heavily at the opening, which was down 2~. For the rest of the day they engaged in "in and ou~" trading, quickly offsetting their purchases by sales. (b) A,co Corp.--For the day as a whole floor traders had total sales of 4,300 shares, representing only 5.3 percent of all volume. However, the timing of these sa]es gave them greater weight than the number of shares would indicate. For example, floor traders sold 2,400 shares from 12:30 to 1:20 when the stock declined from 20~ to 20, the stock’s lowest point in the day until then. Floor traders sold heavily again at about 2:52, selling 1,000 shares in 2 blocks, following which the stock dropped abruptly from 191/s to (c) Bm~wie]¢ Corp.--This was the only one of the eight stocks in which the floor traders had more than twice as many purchases as sales ~or the day. Most of their purchases were concentrated early and late in the day. Two relatively large blocks were acquired at 3:25 when the stock experienced a short but sharp decline. Floor traders assisted the specialist on May 29 at the opening, which was down one point from the previous close. Of 50,000 shares sold, floor traders took 5,000 and the specialist 9,400 shares. (d) International Business Machines Corp.--The stock opened at 397, its high for the day, hit an intraday low of 355 and closed at 361. For the day as a whole floor traders had a sale balance of 1,200 shares, representing 2.1 percent of total volume in the stock. Between about 2 and 3 p.m. when IBM registered its most significant drop, floor traders sold 1,900 shares on balance. By comparison, the specialis~ had net sales of 100 shares, odddot dealers net sales o~ 800 shares, members off-floor net sales of 200 shares, and nonmembersnet pur- chases of 3,000 shares. Toward the end of the hour, a drop of 9 points, from about 374 to near 365, occurred within a few minutes. During this short period the floor traders sold 800 shares and chased none, the specialist bought 700 shares and sold none, and the public bought 1,900 shares and sold 1,800. Thus, at this critica] juncture, floor traders were the only sellers on balance. (e) Standard Oil Co. (New Jersey).--On May 28, floor traders purchased a total of 10,700 shares of this stock and sold 7,300 for purchase balance of 3,~00. Most of their urchases were made in the first 4 hours o~ trading and their sales inPthe next hour after which the stock declined rapidly. During a gradual decline in the stock from 50 at the opening to about 49V2 at 12:30, floor traders bought 5,300 shares and sold 700 shares. Immediately thereafter, the stock ~ell sharply to 48~ and floor traders bought 2,600 shares at or near this low. The stock then stabilized and floor traders made a few small sales and virtually no purchases. They participated actively again after another sharp decline to 475~ during a short time interval just before 2:00 p.m. when they purchased 1,300 shares again a.¢~ or near the low point. As the stock stabilized between 2:00 and 2:50 854 REPORT OF SPECIAL STUDY OF SECURITIES ]~IARKETS p.m., floor traders liquidated their positions to the extent of 4,100 shares. At about 2:50, the stock started to sell off again and experi- enced its largest and most prolonged decline of the day, dropping to 45~. Floor traders continued their selling through the first part. of this drop, selling 1,000 shares, but then withdrew except for ’2 traders who bought a total of 500 shares around the low for the day. d. Members off-~oor A memberof the Exchange is permitted to buy and sell stock under substantially the same conditions as nonmembers, if the order is originated off the floor. Membersand memberfirms may trade off the floor for their personal accounts or for their firm trading accounts. The major limitation imposed by the Exchange upon their trading requires that customers’ market orders must be executed before the members’ own orders. 49 A member who is not on the floor does not have the advantage of being able to observe and evaluate the market as does the specialist, floor trader, or odd-lot dealer. However,he can appraise the market to a better extent than the general public because he has knowledgeo.f the flow of customers’ orders through his firm and may have direct contact with a floor partner. Whether or not any member, other than a specialist,, has a responsibility to maintain an orderly market, or at least to r~frain from contributing towards in- stability, is not considered here2° The following discussion confines itself to a quantitative presentation of the trading by membersoff the floor in the period leading to the market break and during the break itself. (1) Aggregate activity On a weekly basis, members off-floor usually account for about 5 percent of all volume. They generally sell shares on balance on the 1~¥SE, apparently because they acquire shares off-the-board through such means as stock splits and dividends, arbitrage, conversions, and other off-board transactions. Despite their liquidation of off-board acquisitions, they had small weekly purchase balances from September through l~ovember 1961. In December~ members off-floor shifted their policy and began to sell heavily. Beginning with the week ended December1, 1961, through May25, 1962, they sold 4,9¢7~000 shares on balance. Their large net sales contrast with net sales of 205,000 shares for specialists and 538~000shares for floor traders for the period. For the week ending May25, 1962, membersoff-floor purchased on balance for the first time in 26 weeks. The buying which began in the previous week continued into the week of the break. On May 28, members off-floor bought 416,400 shares and sold 271,900 shares, resulting in a purchase balanc~ of 144,- 500 shares. In the heavy trading volume that day they accounted for only 3.5 percent of total round-lot purchases and sales compared with their usual 5 percent. On May 29 they had a very small purchase balance of 7,300 shares and on May 31 they sold 328,900 shares net. Their activity on these 2 latter days was also below normal on a relative basis.

*~ NYSE r~le 92. The rule also prohibits members’ buying at or below, or selling at or above, the price of a customer’s unexecuted limit order to buy or sell, respectively. ~o Congressional expressions of opinion at the time of the enactment of the Exchange Act on the duties and responsibilities of off-floor members are presented in ch. VI. REPORT OF SPECIAL STUDY OF SECURITIES MARKETS 855

(2) Activity in the eight stocks On May 28, members off-floor had net purchases in seven of the eight stocks studied. However, in four of these neither their pur- chases nor sales amounted to us much as 5 percent of total voluIne. The other three stocks and the one stock in which they sold on balance, Korvette, are discussed below. (a) American Telephone c~ Telegraph Co.--Members off-floor had a large purchase balance of 27,100 shares o.n May 28, representing total purchases and sales of 12.6 and 3.0 percent, respectively, of total volume. The heaviest purchases were made between 10:00 and 11:00 when the price was relatively stable; between 12:00 and 1:00 when the price dropped from approximately 109 to 106; and from about 2:15 to 2:45 when the price fell more than 3 points. On May 29 and 31, membersoff-floor sold on balance. (b) Avco Corp.--Members off-floor ~were also heavy p.urchusers in Chic stock on May 28. Th%vpurchased 23,400 shares and sold 3,600, of which 3,100 were short sales. Their pure’hast accounted for 29 per- cent of to, el v.olume, and they were the most a.etive merfrber group in .the sVock. ’This activity was unusual becanse, in the 16 days stud~ied before and a~t.er t~he break, they ~accoun.ted on the ’average ~or only 7 percent of all purchases. To a large extent on May 28 members off- floor concentrated their purchases wi’thin or at the end of periods of sharp price declines. (c) General Motors Corp.--Members off-floor bad a p.umhase bal- ance of 1,300 shares in ~this stock on M’ay 28. They pureha’sed proximately 2,~00 shares xt ,the opening and later in ~he day, .as ~;he price declined, ~hey sold on bMance.Their sales for ,tahe day ’amour~ed to slightly more ~han 5 percent of total volume. The sales were not scattered ,throughout the day but appeared on the ta~)e in large blocks For example, they were responsible for a sale of 1,30"0 shares ’~a.t 10 1,000 shares at about 1:45, and 1,000 shares .at 1:52. These la.¢ter sales were printed on the tape .at ~bout 2:50 and 3:15 respectively and may have a~traeted other .sellers. During ~’he last. 40 minu.’tes of bhe trading day the stock experienced its steepest decline. (d) ~. J. Koreette, Inc.~Thi~s was the on’]y one of .the eight stocks in which members off-floor had a sMe balance on May 28. For the day as a whole members off-floor purchased 2,800 shares and sold 4,300. Most of their sales occurred in ’~he last pat* of ~’he d.av Fromat)oroxi- merely ~ :00 un~til 3:00 p.m., ~hey bou.~ht 500 shares and sold 1 100 sh ares ’~s the s~ck dropped Ino price. from~ 39 to 34. In ,~he nex*, few .minu’tes, star, ring at ~:bou*3:00 p.m., members.off-floor sold 1,400 shares ~n rapid succession .~t 34:, ,the low for the day. These sales were in transactions of 500, 500, and 400 shares. e. Odd-lot dealers Transactionsin round1.o~s by odd-’lot .dealers ’are a.t least p,~stially reflection of odd-lot customers’ purchases and sales. Since odd-lot dealers rarely have a perfect matching of buy and sell orders in stock, ~the associ.~e .odd-lot brokers,s~ w:ho are the odd-lot dealers’ rep- I~esen, tatives on .the floor, mustiniti’ate round-]o¢purchases and ’sales to offset *he o.dd-~]ot purchase or sale balances. In effeeting round-lot

See ch. VI.E. $56 REPORT OF SPECIAL STUDY OF SECURITIES MARKETS sales .to offset, odd-lot brokers are exempt t~rom the Exchange’s floor trading rules. The odd-lot brokers are entrusted with some discretion in initiating round-lot transactions, but the odd-lot dealers have some l’imi’~ on ~the size .of Vhepositions maintained’by .the brokers. Odd-lot dealers sometimesm’a.in~ta’in substantial p.osi~ions in active stocks. In declining marke’ts the dealers o~.ten tend to reduce ’long positions and increase short pos:itions; in rising markots they tend ~o increas~ their invel~tories generally, but in both cases, there is variety amongindividual sto.cl~s. The odd-lot dealer may not offset h,is cus- tomer"sn.et activity ei,ther a.s ’i~t occurs .during .the day or t}or the entire day, .ar~d to ~he extent o~ this "~vi.thholding," ~- h~ {s diverting the ira- p.aot of the odd-lot customer’s curren’~ net activity from Vhe exchange market. (1) Aggregate activity In the last 3 months of 1961, odd-lot customers sold on balance but in the first 5 months of 1962, except for a 7-week period in February and March and the week ended May 25, they had weekly purchase balances. Odd-lot dealers’ round-lot offsets generally ~ollowed this same pattern. However, odd-lot dealers, on a net basis, purchased a smaller numberof shares in round lots than they sold to odd-lot customers (table XIII- 19). Thus, from the first weeek in January 1962, through May 25, they reduced their positions by 226,000 shares. This indicates that odd-lot dealers, like all other membergroups, took a bearish view of the market in early 1962. Odd-lot dealers acco .tinted for a daily average o~ 3.2 percent of round-lot volume in the first 5 months of 1962. On May 28 and 29 they increased their share activity but were responsible for only 2.6 and 2.3 percent of round-lot volume. In part, this modest participa- tion rate may be explained by the fact that odd-lot customers did not have unusually large purchase or sale balances so that the dealers’ offsetting trades were correspondingly small. On May 28, odd-lot dealers bought and sold round lots of 128,210 and 375,520 shares, respectively, for a sale balance of 2¢7,310 shares; odd-lot customers had net sales of 266,812 shares. As .a result, there- fore, odd-lot dealers increased their positions by 19,502 shares. One of the two primary odd-lot dealers decreased its position by 10,228 shares, reflecting an increase in its short position partially offset by a smaller increase in its long position, while the other odd-lot dealer had a net increase in position of 3~:,0¢2 shares for the day as a result of a substantial increase in long position. (The small odd-lot dealers had an aggregate decrease in position of ~,312 shares.) The odd-lot dealers increased their aggregate positions on the 28th by some 19,000 shares and to this extent they insulated the round-lot market ~rom their customers net sales activity. It is difficult to evalu- ate the net effect of this withholding, however, in terms o~ its effect on the market as this depends on their activity in particular stocks at particu]ar times.

~Withholding, the opposite of perfect offset, occurs when an odd-lot dealer delays transmittal of his customer’s net activity to the ~o.undAot market. An odd-lot dealer ~nay also sell more than hts customers sell (~r buy more than they buy, thus reducing or a44ing to his inventory, perhaps in anticipation of odd-lot customers’ activity. REPORT OF SPECIAL STUDY OF SECURITIES MARKETS 857 (2) Activity in the eightstoclcs In all eight of the stocks studied~ the odd-lot dealers~ trading in round lots~ purchased less on a percentage basis on May28 than they did on the average in the 16 nonbreak days; but in five of the eight stocks they sold more oa a percentage basis that day than on the other days. The two principal odd-lot dealers combineddecreased their inventories or added to their short positions on May28~ 1969,, in six of the eight stocks studied. This means that in six of these stocks they sold more shares in the round-lot market than was necessary to offset odd-lot customers’ sales~ or that they filled odd-lot customers~ purchase orders from their inventory or by selling short. Thus, net purchases were not fully reflected in the round-lot market on this critical day. Ex- amples of this are shown below in a more detailed discussion of trad- ing by odd-lot dealers in the two stocks among the eight which had the largest odd-lot volume on May9,8. (a) American Telephone & Telegraph Co.---On May !~8, all odd-lot purchases and sales represented 9 percent of total NYSEvolume. However, odd-lot purchases and sales of AT & T amount to 9,0 per- cent of volume in the stock. ’This large odd-lot volume is not unusual in a high-priced issue. In the face of this heavy odd-lot customer volume, however, the odd-lot dealers were relatively inactive with their purchases ~epresenting only 9,.0 percent and their sales 0.1 percent o:[ total NYSEround-lot volume. This relative inactivity may be ex- plained primarily by the fact that the odd-lot customers’ substantial purchases and sales were largely in balance, with resulting net pur- chases of 7,700 shares. In addition, odd-lot dealers withheld ~,100 shares of their customers’ purchase balance from the auction market, buying 5,800 shares and selling 9,00 shares. (b) International Business Machines Corp.--Odd-lot dealers bought 9,,300 shares and sold 7,500 shares of this stock in round lots on May9,8. Their net sales of 5,9,00 shares amount to 9.0 percent of all round-lot sales that day and they were the largest net sellers of any member group. Odd-lot customers, who accounted for 33 percent of total volumein the stock, had a sale balance ot~ 4,010 shares for the day. The odd-lot dealers therefore reduced their positions by 1,190 shares. It is of interest to note that customers of 1 of the 9. major firms bought 19_4 shares on balance for the day while that firm sold 800 shares net in round lots. The firm was long 753 shares at the opening so that during the day it not only liquidated its position but established a short position of 171 shares as well.

The detailed analysis of trading in L~dividual stocks reveals the in- adequacy of analysis based exclusively on aggregate data. For vir- tually every generalization made, important exceptions occurred that might have exerted a significant influence on the course of prices. Nonmemberscharacteristically dominated transactions in the stocks studied~ yet in the case of IBM: memberpurchases were more import- ant than those of nonmembersin 8 out of the 16 days covered.

~a All analyses of trading in the eight stocks by odddot dealers are based on data for the two largest odd-lot 4ealers only. These firms accounted, for 99 percent of total volume by odd-lot dealers on ~Iay 28, 1962. REPORT OF SPECIAL STUDY OF SECURITIES MARKETS

Amongnonmembers, public individuals in the aggregate usually were the most imp’ortant component. On May 11, however, the re- porting open-end invest~nent companies constituted 48.5 percent of NYSEvoiume, a higher percentage than that of the public individuals. Then again on various days foreign activities in IBM, usually small in relation to the other participants, represented the largest segment of the transactions reported by the 9,,5.~ firms. Also, during the break, the NYSErep.orted that individuals had a sale balance of 1,572,000 shares on May’28, a sale balance of 1,110,000 shares on May 29, a.nd a purchase balance of 1,946,000 shares on May 31. Illustrative of the inadequacy of such aggregate data is that in only one of the eight stocks did the buying and selling of indi- viduals as reported by the 25 fi~wns result in a duplication of the over- ~11 pattern. A closer examination revealed that with respect to Sts~ndard Oil, IBM, and AT& T, individuals were net sellers on all 3 d~ys. In the case of Brunswick, purchases and sales were virtually in balance on May 28, and there was a purchase balance on May 29 and a slight sale balance on May 31. In Korvette, there ~vas a pur- chase balance on May 28 and 29 followed by a sale balance on May 31. In U.S. Steel there was a slight stock purchase balance on May ~8, a s~le balance on May 29, and a purchase balance once more on May 31. Aveo had a sale balance on May ~8, followed by a purchase balance on May ~9 and May 31. Only in the ease of General Motors did individuals repeat their overall performance of selling on balance on May 28 and ~9, while buying on balance on May 31. The aggregate analysis indicated that institutions tend to accumu- late stock on s~ net basis. In every month from September 1961 through June 1962, the institutions studied had a purchase balance which was particularly large beginni.ng March 196’2. Moreover, on an overall basis, institutions were nnportant net buyers during the break. Against this general picture may be matched the institutions’ persistent selling of U.S. Steel from January through June 1962, and their heavy net sales of IBM from September 1961 through January 196~. Similarly, even though the open-end investment companies were net buye,r,s during the period of the market break and particu- larly in the ~eek before, their persistent selling of Brunswick and U.S. Steel from the fall of 1961 through June 1962 may have con- tributed to the weakness in those stocks. The substantial selling bal- ances of the open-end investment companies in General Motors occurred during most of Mayas the price of the stock was declining. These were stocks in the public eye and the selling pressures on them during critical times may well have exerted a significant influence on the overall market. With respect to members, the story is the same. In the aggregate, their transactions represented 20 to. 24 percent of total volume but on some day or days in each of the eights sf:oeks they accounted for at least 35 percent of all purchases or sales on the NYSE,and in several of the stocks the percentage reached 50 percent or more on various days. While specialists had a purchase balance on May 9~8, they were net sellers of IBM, Korvette, and U.S. Steel. Furthermore, analysis of their intraday tradin.g in these stocks reveals various instances in which they were p.asmve or were actually net sellers at critical junctures. REPORT OF SPECIAL STUDY OF SECURITIES MARKETS $59

E. SUil~VIARY, CONCLUSIONS, AND RECO:~II~[ENDATIONS Shortly after the market break of May P~8-31 the Special Study was asked to add to its agenda an examination of that important occurrence. Some of the results of its inquiry are reflected in other chapters (especially VI and VII) and a general summary is con- rained in this chapter. In view of the fact that the :NYSEhas pub- lished its own study 54 containing relevant aggregated data for the 3 particular days, the Special Study has sought to avoid duplication of that analysis. Instead it has attempted to take a somewhat wider look, by examining trading on 16 additional days, and at the same time a closer look, by studying specific stocks and disaggregated data. In its analysis of the disaggregated data the study fom~dthat while there were general patterns of behavior, there were also striking departures from the overall picture. For example, odd-lot customers in the aggregate were net sellers on May28, but they had a purchase balance in AT & T. The open-end investment companies studied, on the other hand, were overall net buyers o.f stocl~s on that day but were sellers on balance in General Motors and U.S. Steel and had no transactions in Avco or Brunswick. Similarly, although spe- cialists as a group, had a purchase balance, they were relative.ly large net sellers of Korvette and had modest sale balances in IBM and U.S. Steel. These variations in the practices of the partici- pants in individual issues reveal the inadequacy of aggregated data alone to portray realistically the diversity of members’ and non- members’transactions in individual stocks. Neither this study nor that of the :New York Stock Exchange was able to isolate and identify the "causes" of the market events of May 28, 29, and 31. There was some speculation at the time that these events might be the result of some conspiracy or deliberate misconduct. Upon the basis of the study’s inquiry, there is no evidence whatsoever that the break was deliberately precipitated by any person or group .or that there was any manipulation or illegal conduct in the function- ~ng of the market. The avalanche of orders which came into the market during this period subjected the market mechanisms to extraordinary strain, and m many respects they did not~ function in a normal way. Particularly significant were the lateness of the tape and the consequent inability of investors to predict accurately the prices at which market orders would be executed. Further indicative of the disruption of the trading mechanisms, some odd-lot orders on May 28 were not executed at the first round-lot sale following receipt as required, but at the day’s closing price, in most instances considerably lower, plus or minus the odd-lot differential.

~’ NYSE, "The Stock ~Iarket Under Stress," (1963). $60 REPORT OF SPECIAL STUDY OF SECURITIES 1VfARKETS

On the 3 days of the m~rket break the percentage distribution for purchases and sales by types of orders ~ on the NYSEwas as follows"

TABLS XIII-d.--Distribution of types of arders on the NYSE, May 28, 29, and 3.I [Percent]

Market Limit Stop

May28, tot~l ...... 53.1 42.4 4.5 Purchases ...... 35.4 64.1 .5 Sales ...... 70.1 21.5 8.4 May29, total ...... 69.3 29.1 1.6 Purchases ...... 64.5 34.6 .9 Sales ...... 74.1 23.5 2.4 May31, total ...... 60.5 38.5 1.0 Purchases ...... 68.2 31.3 .5 Sales ...... 51.3 47.2 1.5

Source: NYSE,"The Stock Market UnderStress," p. 37 (1963). It is noteworthythat on M~y9,8, 70.1 percent of public sell orders were m~rket orders and another 8.4 percent were stop orders, whereas 9,1.5 percent .were limit orders. Onthe buy side, on the other hand, 64.1 percent were limit orders, and 35.9 percent market and stop orders. Since May29 was characterized by ~ continuation of the sharp decline during the earlier part of the d~y and ~ very sharp recovery in the l~ter part of the d~y, ~nd since it was not possible to allocate orders between these two phases, significant relationships in terms of types of orders could not be established. On May31, how- ever, whenprices movedsharply upward,there was a distinct reversal of the p~ttern from that of Monday:on the purchase side 68.7 per- cent of orders were market and stop orders, whereas on the sell side only 59,.8 percent were marketand stop orders. The relatively large volume of sell-stop orders on May9,8 is also worthy of note. As already mentioned, such an order is used as ~ protective measur~to assure ~ promptsale if the m~rketprice reaches or f~lls below ~ previously specified figure, and it becomes a market order when that price is reached. Thus, ~ sharp decline such as that of May28, already involving a heavy preponderance of market sell orders as compared.with buy orders, produces ~ separate source of market orders as stop orders are triggered by the decline. The sell- stop orders hem by specialists on May9,8 may not have been entered on that day; some may well have been placed at any time previously and h~ve comeinto play as prices fell. The "snowballing" effect o.f stop orders on May28 was pointed out by a specialist whotestified" * * * the book was heavy with stop orders, and they, as much as anything, were responsible for the decline, with an overhanging volume of market short

r~q?hree main types of order,s are used to buy or sell securities in the auction market o~ the N~YSE: market orders, limit order.s, and stop orders. Briefly, a market order is one to buy or sell at the best price avail~ble. A limit order is one to buy or sell at a specific price or better; on the sell side the specitied price would be above the prevailing market, and on the buy si~e, below the prevailing market. A stop order, sometimes called a "stop loss" order, specifies a price at which, if the market moves adversely, the cus- tomer desires an execution. Iaf the order is on the sell side, it specifies a price below that prevailing ; on the buy side, price above that prevailing~the reverse of the situation in limit orders~ It does not, however, guarantee execution at the specified price, but merely becomes a market order if and when that price is reached. It is to be expected that the volume of sell-stop orders would ordinarily exceed that of buy-stop orders. REPORT OF SPECIAL STUDY OF SECURITIES MARKETS

orders. The bid had to be dropped considerably to take care of the new stop orders that were put into effect * * * The volume of short selling in the aggregate, and for certain in- dividual stocks, by classes of participants, is shown elsewhere in the report, 5G but these figures do not necessarily reveal the full impact of short selling. In testimony taken by the study, specialists indi- cated that there was a significant amount of potential short selling (brokers in tt~e "crowd" waiting for an uptick) which was never real- ized in transactions. This potential short selling overhanging the market may well have prompted some specialists to moderate their stabilizing activities, since they would know that any rally would be met by short-sale orders in the "crowd." As one specialist put it, short selling during the break acted to "lengthen the time that it took a stock to go up because there had to be substantially more buyers to movethe stock up. * * *" The Exchange A~t makes it clear that there is an important public interest in the effects of rapid price fluctuations both up and down. The Act states as one of the reasons for its passage the fact that "na- tional emergencies * * * are precipitated, intensified, and prolonged by * * * sudden and unreasonable fluctuations of security prices and b.y excessive speculation. * * *" ~7 Accordingly, the Commission is given the authority and responsibility-- if in its opinion the public interest so requires, summarily to suspend trading in any registered security on any national securities exchange for a period not exceeding 10 days, or with the approval of the President, summarily to suspend all trading on any national securities exchange for a period not exceeding 90 days. The power to suspend all trading, on an exchange is indeed an awe- some one. as indicated by the reqmrement of Presidential approval, and the (~ommission has never invoked it. Once market changes be- came so chaotic as to warrant halting all trading on the exchanges, it is possible that investor tensions would be so acute that unexpected and severe reactions might follo~v from the suspension itself. On the other hand, assuming that any intermediate, technical meas- ures-i.e., measures short of suspension of all trading--would be feasible and desirable, it obviously is not practicable to wait until a severe break is in progress to determine what they may be. The un- certainties and pressures existing under such conditions militate against the development of a sound course of action. Nor is it possible at the time of a market break, unless arrangements for gathering in- formation have been worked out in advance, to obtain speedily the kind of current and meaningful trading data which the Commission and other Government agencies might consider useful in discharging their responsibilities. Yet, once a break has passed, there is a tendency to forget the concerns existing at the time and the apprehensions as to what might happen should it continue. The history of the May28 market break reveals that a complex inter- action of causes and effects--including rational and emotional moti- vations as well as a variety of mechanisms and pressures--may sud- .denly create a downward spiral of great velocity and force. This, in turn, may change the impact of various normal market mechanisms,

Ch. VI.H.5.b. Exchange Act, sec. 2(4). 862 REPORT OF SPECIAL STUDY OF SECURITIES 1VIARKETS and thus temporarily impair the market’s fair and orderly character. Where the latter situation prevails, a public interest in orderly mar- kets, quite distinguishable from any public intervention in the setting of price levels, maycome into play. The question thus arises whether it would be desirable and feasible for the Commission and the industry jointly to formulate programs for exchanging information and/or for the taking of intermediate, technical steps--short of suspension of trading--that would be de- signed to provide market conditions as orderly as possible in a period of .stress, even though they could not, of course, be expected to alter major market trends. The Special Study is of the view that, whether or not such programs would ultimately be found practicable or de- sirable, the question is one deserving further exploration. 5s Any pro- gram of intermediate measures that might be evolved would presum- ably contemplate action to be taken primarily by the industry as dis- tinguished from the Commission, which would remain essentially in the role of overseer of self-regulatory action. It would be unrealistic and indeed illusory to believe that the narrow and technical pew. ers possessed by the Commissionitself could ever prevent basic price changes. The Commission’s role is primarily regulatory, not economic. Traditionally and consistently, it has ex- ercised its powers in such manner as to avoid dealing with price levels or permitting any misconception that it was dealing with price levels. Nothing in this chapter is intended to suggest a change in this role in the direction of "managing" price movements or purposefully af- fecting prices. The NYSEis already endeavoring to develop improved equipment which should greatly ameliorate the problems arising from tape late- ness. The implementation of various specific recommendations made elsewhere in the report, in part upon the basis of data relating to the market break, with respect to such matters as short selling, the capital position of specialists, floor trading and odd-lot transactions, should also tend to improve the ability of Exchange mechanisms to ~unction more effectively in times of stress. The study being made by the Division of Trading and Exchanges with respect to stop orders should contribute to this effort. The Special Study concludes and recommends: Neither the Study nor the NYSEhas been able to ascertain the precipitating "causes" of the May 1962 market break. However, analysis of disaggregated market data has permitted identifica- tion of certain specific factors in the operation of market mech- anisms that may have accentuated its severity. At most, any

~s After war was declared in September 1939, lines of direct communication to important sources of information in the financial community were established and liaison with the national securities exchanges was developed : "Through its machinery for gathering as much information as possible, it kept constant scrutiny over the volume and trend of orders as they came into the leading brokerage offices before those orders reached the floor. Each morning before the markets opened the Commission anc~ its experts were in contact with its. sources of information to find out the character an4 size of the brokerage orders which had accumulated overnight. It kept track of the effect of market changes upon margin accounts. It received current reports on the size of short positions and the condition of the books of the specialists in various leading stocks on the floors of the varlou.s exchanges. It was able, in cooperation with the New York Stock Exchange, the r~reasury, and certain houses special- izing in foreign dealings, to judge the trend and volume of foreign transactions." S.E.C. Sixth Annual Report, p. 89 (1940). Similar steps were taken o,n later occasions such as at the ti,me of President Eisen- hower’s heart attack. REPORT OF SPECIAL STUDY OF SECURITIES I~ARKETS 863 measures that might be taken with reference to such factors could only be addressed to ameliorating their impact. The Commis- sion’s role is to promote an orderly market and not to affect fun- damental economic forces or price trends. The following recom- mendation must be viewed in this context. The Commission and representatives of the industry, particu- larly the exchanges, should make a joint study of possible inter- mediate measures, short of suspending trading, that might be invoked to assure minimum disruption of the fair and orderly functioning of the securities markets in times of severe market stress. While the Special Study has not undertaken to evaluate the possibilities, the types of intermediate measures to be con- sidered might include such things as limitations on short selling (see ch. VI.tt, recommendation 3), special provisions in respect the handling of stop sell orders or market sell orders, and tem- porary interruption of trading in individual securities under predefined circumstances. It is possible that the implications of such actions could be tested in advance through the t~se of simu- lation techniques on a computer. There should also be Commis- sion-industry consultation with a view to collecting certain crucial types of trading information that might be helpful in connection with possible application of any of such intermediate measures or that might be useful in times of market stress to other govern- mental agencies having wider economic responsibilities. ChartXlll-I { Part I J AMERICAN TELEPHONE& TELEGRAPHCO. (PARTII continuedon next PRICE MOVEMENT, VOLUMEAND MEMBERS’ TRANSACTIONS MAY 28, 1962